{
  "best_global_step": 38160,
  "best_metric": 0.4749854803085327,
  "best_model_checkpoint": "saves_multiple/lntuning/llama-3-8b-instruct/train_codealpacapy_123_1762561812/checkpoint-38160",
  "epoch": 20.0,
  "eval_steps": 1908,
  "global_step": 38160,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002620545073375262,
      "grad_norm": 1.8199111223220825,
      "learning_rate": 5.241090146750525e-08,
      "loss": 1.257,
      "num_input_tokens_seen": 2944,
      "step": 5
    },
    {
      "epoch": 0.005241090146750524,
      "grad_norm": 0.8825773000717163,
      "learning_rate": 1.1792452830188679e-07,
      "loss": 1.1154,
      "num_input_tokens_seen": 6816,
      "step": 10
    },
    {
      "epoch": 0.007861635220125786,
      "grad_norm": 1.177601933479309,
      "learning_rate": 1.8343815513626837e-07,
      "loss": 1.2213,
      "num_input_tokens_seen": 9760,
      "step": 15
    },
    {
      "epoch": 0.010482180293501049,
      "grad_norm": 0.7590242624282837,
      "learning_rate": 2.489517819706499e-07,
      "loss": 1.1648,
      "num_input_tokens_seen": 13472,
      "step": 20
    },
    {
      "epoch": 0.01310272536687631,
      "grad_norm": 2.1295323371887207,
      "learning_rate": 3.144654088050315e-07,
      "loss": 1.2346,
      "num_input_tokens_seen": 16864,
      "step": 25
    },
    {
      "epoch": 0.015723270440251572,
      "grad_norm": 1.1735913753509521,
      "learning_rate": 3.79979035639413e-07,
      "loss": 1.3564,
      "num_input_tokens_seen": 20352,
      "step": 30
    },
    {
      "epoch": 0.018343815513626835,
      "grad_norm": 1.3178958892822266,
      "learning_rate": 4.454926624737945e-07,
      "loss": 1.3339,
      "num_input_tokens_seen": 23008,
      "step": 35
    },
    {
      "epoch": 0.020964360587002098,
      "grad_norm": 1.3185662031173706,
      "learning_rate": 5.110062893081761e-07,
      "loss": 1.0737,
      "num_input_tokens_seen": 25920,
      "step": 40
    },
    {
      "epoch": 0.02358490566037736,
      "grad_norm": 1.7813935279846191,
      "learning_rate": 5.765199161425577e-07,
      "loss": 1.2294,
      "num_input_tokens_seen": 28672,
      "step": 45
    },
    {
      "epoch": 0.02620545073375262,
      "grad_norm": 0.9910851716995239,
      "learning_rate": 6.420335429769392e-07,
      "loss": 1.0775,
      "num_input_tokens_seen": 31488,
      "step": 50
    },
    {
      "epoch": 0.028825995807127882,
      "grad_norm": 1.1450971364974976,
      "learning_rate": 7.075471698113208e-07,
      "loss": 1.2133,
      "num_input_tokens_seen": 34496,
      "step": 55
    },
    {
      "epoch": 0.031446540880503145,
      "grad_norm": 0.7359253168106079,
      "learning_rate": 7.730607966457024e-07,
      "loss": 1.2943,
      "num_input_tokens_seen": 38112,
      "step": 60
    },
    {
      "epoch": 0.034067085953878404,
      "grad_norm": 1.3049163818359375,
      "learning_rate": 8.38574423480084e-07,
      "loss": 1.3676,
      "num_input_tokens_seen": 40864,
      "step": 65
    },
    {
      "epoch": 0.03668763102725367,
      "grad_norm": 0.6989263296127319,
      "learning_rate": 9.040880503144654e-07,
      "loss": 1.2801,
      "num_input_tokens_seen": 44192,
      "step": 70
    },
    {
      "epoch": 0.03930817610062893,
      "grad_norm": 1.739898920059204,
      "learning_rate": 9.69601677148847e-07,
      "loss": 1.4051,
      "num_input_tokens_seen": 46784,
      "step": 75
    },
    {
      "epoch": 0.041928721174004195,
      "grad_norm": 3.329134464263916,
      "learning_rate": 1.0351153039832286e-06,
      "loss": 1.3892,
      "num_input_tokens_seen": 50176,
      "step": 80
    },
    {
      "epoch": 0.044549266247379454,
      "grad_norm": 0.8248217701911926,
      "learning_rate": 1.10062893081761e-06,
      "loss": 1.0898,
      "num_input_tokens_seen": 54400,
      "step": 85
    },
    {
      "epoch": 0.04716981132075472,
      "grad_norm": 1.699541687965393,
      "learning_rate": 1.1661425576519918e-06,
      "loss": 1.4386,
      "num_input_tokens_seen": 56992,
      "step": 90
    },
    {
      "epoch": 0.04979035639412998,
      "grad_norm": 1.5277692079544067,
      "learning_rate": 1.2316561844863732e-06,
      "loss": 1.0163,
      "num_input_tokens_seen": 60896,
      "step": 95
    },
    {
      "epoch": 0.05241090146750524,
      "grad_norm": 0.7043004035949707,
      "learning_rate": 1.2971698113207548e-06,
      "loss": 1.1498,
      "num_input_tokens_seen": 64576,
      "step": 100
    },
    {
      "epoch": 0.055031446540880505,
      "grad_norm": 2.384634017944336,
      "learning_rate": 1.3626834381551364e-06,
      "loss": 1.3792,
      "num_input_tokens_seen": 67616,
      "step": 105
    },
    {
      "epoch": 0.057651991614255764,
      "grad_norm": 0.9512035250663757,
      "learning_rate": 1.4281970649895177e-06,
      "loss": 1.0257,
      "num_input_tokens_seen": 72544,
      "step": 110
    },
    {
      "epoch": 0.06027253668763103,
      "grad_norm": 1.112167477607727,
      "learning_rate": 1.4937106918238995e-06,
      "loss": 1.2987,
      "num_input_tokens_seen": 75648,
      "step": 115
    },
    {
      "epoch": 0.06289308176100629,
      "grad_norm": 1.4007397890090942,
      "learning_rate": 1.559224318658281e-06,
      "loss": 1.386,
      "num_input_tokens_seen": 78368,
      "step": 120
    },
    {
      "epoch": 0.06551362683438156,
      "grad_norm": 1.4612343311309814,
      "learning_rate": 1.6247379454926625e-06,
      "loss": 1.239,
      "num_input_tokens_seen": 81952,
      "step": 125
    },
    {
      "epoch": 0.06813417190775681,
      "grad_norm": 1.238466501235962,
      "learning_rate": 1.690251572327044e-06,
      "loss": 1.1895,
      "num_input_tokens_seen": 84960,
      "step": 130
    },
    {
      "epoch": 0.07075471698113207,
      "grad_norm": 1.7679224014282227,
      "learning_rate": 1.7557651991614257e-06,
      "loss": 1.3062,
      "num_input_tokens_seen": 88576,
      "step": 135
    },
    {
      "epoch": 0.07337526205450734,
      "grad_norm": 1.309597134590149,
      "learning_rate": 1.821278825995807e-06,
      "loss": 1.0192,
      "num_input_tokens_seen": 92160,
      "step": 140
    },
    {
      "epoch": 0.0759958071278826,
      "grad_norm": 0.9449553489685059,
      "learning_rate": 1.8867924528301887e-06,
      "loss": 1.1054,
      "num_input_tokens_seen": 95840,
      "step": 145
    },
    {
      "epoch": 0.07861635220125786,
      "grad_norm": 1.639782190322876,
      "learning_rate": 1.9523060796645703e-06,
      "loss": 1.1889,
      "num_input_tokens_seen": 98816,
      "step": 150
    },
    {
      "epoch": 0.08123689727463312,
      "grad_norm": 0.8539170622825623,
      "learning_rate": 2.017819706498952e-06,
      "loss": 0.9874,
      "num_input_tokens_seen": 102880,
      "step": 155
    },
    {
      "epoch": 0.08385744234800839,
      "grad_norm": 1.5906200408935547,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 1.1657,
      "num_input_tokens_seen": 105920,
      "step": 160
    },
    {
      "epoch": 0.08647798742138364,
      "grad_norm": 0.8684478402137756,
      "learning_rate": 2.148846960167715e-06,
      "loss": 1.3796,
      "num_input_tokens_seen": 108800,
      "step": 165
    },
    {
      "epoch": 0.08909853249475891,
      "grad_norm": 1.1275465488433838,
      "learning_rate": 2.2143605870020966e-06,
      "loss": 1.088,
      "num_input_tokens_seen": 111968,
      "step": 170
    },
    {
      "epoch": 0.09171907756813417,
      "grad_norm": 1.462103247642517,
      "learning_rate": 2.279874213836478e-06,
      "loss": 1.1693,
      "num_input_tokens_seen": 114976,
      "step": 175
    },
    {
      "epoch": 0.09433962264150944,
      "grad_norm": 3.6767776012420654,
      "learning_rate": 2.34538784067086e-06,
      "loss": 1.4241,
      "num_input_tokens_seen": 117568,
      "step": 180
    },
    {
      "epoch": 0.09696016771488469,
      "grad_norm": 1.29606032371521,
      "learning_rate": 2.410901467505241e-06,
      "loss": 1.0373,
      "num_input_tokens_seen": 120384,
      "step": 185
    },
    {
      "epoch": 0.09958071278825996,
      "grad_norm": 0.7676098942756653,
      "learning_rate": 2.476415094339623e-06,
      "loss": 1.0023,
      "num_input_tokens_seen": 123680,
      "step": 190
    },
    {
      "epoch": 0.10220125786163523,
      "grad_norm": 1.2269623279571533,
      "learning_rate": 2.5419287211740044e-06,
      "loss": 1.2404,
      "num_input_tokens_seen": 127168,
      "step": 195
    },
    {
      "epoch": 0.10482180293501048,
      "grad_norm": 1.3517199754714966,
      "learning_rate": 2.6074423480083857e-06,
      "loss": 0.9805,
      "num_input_tokens_seen": 130176,
      "step": 200
    },
    {
      "epoch": 0.10744234800838574,
      "grad_norm": 1.1431916952133179,
      "learning_rate": 2.6729559748427675e-06,
      "loss": 1.1461,
      "num_input_tokens_seen": 133056,
      "step": 205
    },
    {
      "epoch": 0.11006289308176101,
      "grad_norm": 1.1254953145980835,
      "learning_rate": 2.738469601677149e-06,
      "loss": 1.0589,
      "num_input_tokens_seen": 136864,
      "step": 210
    },
    {
      "epoch": 0.11268343815513626,
      "grad_norm": 1.3840417861938477,
      "learning_rate": 2.8039832285115303e-06,
      "loss": 1.4862,
      "num_input_tokens_seen": 140064,
      "step": 215
    },
    {
      "epoch": 0.11530398322851153,
      "grad_norm": 1.384464144706726,
      "learning_rate": 2.869496855345912e-06,
      "loss": 1.054,
      "num_input_tokens_seen": 143712,
      "step": 220
    },
    {
      "epoch": 0.1179245283018868,
      "grad_norm": 1.3769123554229736,
      "learning_rate": 2.935010482180294e-06,
      "loss": 1.1625,
      "num_input_tokens_seen": 146880,
      "step": 225
    },
    {
      "epoch": 0.12054507337526206,
      "grad_norm": 1.1505098342895508,
      "learning_rate": 3.0005241090146753e-06,
      "loss": 1.0283,
      "num_input_tokens_seen": 150560,
      "step": 230
    },
    {
      "epoch": 0.12316561844863731,
      "grad_norm": 1.0532952547073364,
      "learning_rate": 3.0660377358490567e-06,
      "loss": 1.2707,
      "num_input_tokens_seen": 153664,
      "step": 235
    },
    {
      "epoch": 0.12578616352201258,
      "grad_norm": 1.7744472026824951,
      "learning_rate": 3.131551362683438e-06,
      "loss": 0.9751,
      "num_input_tokens_seen": 157568,
      "step": 240
    },
    {
      "epoch": 0.12840670859538783,
      "grad_norm": 1.3453725576400757,
      "learning_rate": 3.1970649895178203e-06,
      "loss": 1.0809,
      "num_input_tokens_seen": 160512,
      "step": 245
    },
    {
      "epoch": 0.1310272536687631,
      "grad_norm": 0.5458435416221619,
      "learning_rate": 3.2625786163522012e-06,
      "loss": 0.9947,
      "num_input_tokens_seen": 164512,
      "step": 250
    },
    {
      "epoch": 0.13364779874213836,
      "grad_norm": 2.148913860321045,
      "learning_rate": 3.3280922431865826e-06,
      "loss": 1.2133,
      "num_input_tokens_seen": 167200,
      "step": 255
    },
    {
      "epoch": 0.13626834381551362,
      "grad_norm": 1.9223712682724,
      "learning_rate": 3.393605870020965e-06,
      "loss": 1.2308,
      "num_input_tokens_seen": 169984,
      "step": 260
    },
    {
      "epoch": 0.1388888888888889,
      "grad_norm": 1.0874276161193848,
      "learning_rate": 3.459119496855346e-06,
      "loss": 1.0836,
      "num_input_tokens_seen": 173856,
      "step": 265
    },
    {
      "epoch": 0.14150943396226415,
      "grad_norm": 1.5551851987838745,
      "learning_rate": 3.524633123689727e-06,
      "loss": 1.4256,
      "num_input_tokens_seen": 176192,
      "step": 270
    },
    {
      "epoch": 0.1441299790356394,
      "grad_norm": 0.7500717639923096,
      "learning_rate": 3.5901467505241094e-06,
      "loss": 0.8757,
      "num_input_tokens_seen": 179840,
      "step": 275
    },
    {
      "epoch": 0.14675052410901468,
      "grad_norm": 1.200981855392456,
      "learning_rate": 3.6556603773584908e-06,
      "loss": 1.0033,
      "num_input_tokens_seen": 182976,
      "step": 280
    },
    {
      "epoch": 0.14937106918238993,
      "grad_norm": 1.1390191316604614,
      "learning_rate": 3.7211740041928726e-06,
      "loss": 1.0193,
      "num_input_tokens_seen": 186400,
      "step": 285
    },
    {
      "epoch": 0.1519916142557652,
      "grad_norm": 1.207925796508789,
      "learning_rate": 3.786687631027254e-06,
      "loss": 1.0793,
      "num_input_tokens_seen": 190144,
      "step": 290
    },
    {
      "epoch": 0.15461215932914046,
      "grad_norm": 1.5114232301712036,
      "learning_rate": 3.852201257861635e-06,
      "loss": 1.3219,
      "num_input_tokens_seen": 193344,
      "step": 295
    },
    {
      "epoch": 0.15723270440251572,
      "grad_norm": 1.7358523607254028,
      "learning_rate": 3.917714884696017e-06,
      "loss": 0.9976,
      "num_input_tokens_seen": 195904,
      "step": 300
    },
    {
      "epoch": 0.159853249475891,
      "grad_norm": 1.5196449756622314,
      "learning_rate": 3.9832285115303985e-06,
      "loss": 1.1408,
      "num_input_tokens_seen": 198944,
      "step": 305
    },
    {
      "epoch": 0.16247379454926625,
      "grad_norm": 2.1493585109710693,
      "learning_rate": 4.04874213836478e-06,
      "loss": 1.354,
      "num_input_tokens_seen": 201792,
      "step": 310
    },
    {
      "epoch": 0.1650943396226415,
      "grad_norm": 0.9518304467201233,
      "learning_rate": 4.114255765199162e-06,
      "loss": 0.9747,
      "num_input_tokens_seen": 206080,
      "step": 315
    },
    {
      "epoch": 0.16771488469601678,
      "grad_norm": 1.3510665893554688,
      "learning_rate": 4.1797693920335435e-06,
      "loss": 1.1497,
      "num_input_tokens_seen": 209376,
      "step": 320
    },
    {
      "epoch": 0.17033542976939203,
      "grad_norm": 1.2897526025772095,
      "learning_rate": 4.245283018867925e-06,
      "loss": 1.3004,
      "num_input_tokens_seen": 212160,
      "step": 325
    },
    {
      "epoch": 0.17295597484276728,
      "grad_norm": 0.8458373546600342,
      "learning_rate": 4.310796645702306e-06,
      "loss": 0.9316,
      "num_input_tokens_seen": 216128,
      "step": 330
    },
    {
      "epoch": 0.17557651991614256,
      "grad_norm": 1.216750979423523,
      "learning_rate": 4.376310272536688e-06,
      "loss": 2.0594,
      "num_input_tokens_seen": 219232,
      "step": 335
    },
    {
      "epoch": 0.17819706498951782,
      "grad_norm": 1.0281860828399658,
      "learning_rate": 4.441823899371069e-06,
      "loss": 1.0318,
      "num_input_tokens_seen": 222464,
      "step": 340
    },
    {
      "epoch": 0.18081761006289307,
      "grad_norm": 2.2409844398498535,
      "learning_rate": 4.507337526205451e-06,
      "loss": 1.4151,
      "num_input_tokens_seen": 225184,
      "step": 345
    },
    {
      "epoch": 0.18343815513626835,
      "grad_norm": 1.4074510335922241,
      "learning_rate": 4.572851153039833e-06,
      "loss": 1.2186,
      "num_input_tokens_seen": 228000,
      "step": 350
    },
    {
      "epoch": 0.1860587002096436,
      "grad_norm": 0.8900182843208313,
      "learning_rate": 4.638364779874214e-06,
      "loss": 1.0458,
      "num_input_tokens_seen": 230944,
      "step": 355
    },
    {
      "epoch": 0.18867924528301888,
      "grad_norm": 1.765623927116394,
      "learning_rate": 4.703878406708595e-06,
      "loss": 1.0896,
      "num_input_tokens_seen": 233728,
      "step": 360
    },
    {
      "epoch": 0.19129979035639413,
      "grad_norm": 1.5297547578811646,
      "learning_rate": 4.769392033542977e-06,
      "loss": 1.0326,
      "num_input_tokens_seen": 236480,
      "step": 365
    },
    {
      "epoch": 0.19392033542976939,
      "grad_norm": 2.072105646133423,
      "learning_rate": 4.834905660377359e-06,
      "loss": 1.2213,
      "num_input_tokens_seen": 239488,
      "step": 370
    },
    {
      "epoch": 0.19654088050314467,
      "grad_norm": 1.3542563915252686,
      "learning_rate": 4.90041928721174e-06,
      "loss": 1.2331,
      "num_input_tokens_seen": 242144,
      "step": 375
    },
    {
      "epoch": 0.19916142557651992,
      "grad_norm": 1.1204644441604614,
      "learning_rate": 4.965932914046122e-06,
      "loss": 1.0927,
      "num_input_tokens_seen": 245216,
      "step": 380
    },
    {
      "epoch": 0.20178197064989517,
      "grad_norm": 1.3540819883346558,
      "learning_rate": 5.031446540880504e-06,
      "loss": 1.0174,
      "num_input_tokens_seen": 249632,
      "step": 385
    },
    {
      "epoch": 0.20440251572327045,
      "grad_norm": 0.8183528780937195,
      "learning_rate": 5.096960167714885e-06,
      "loss": 1.1441,
      "num_input_tokens_seen": 253312,
      "step": 390
    },
    {
      "epoch": 0.2070230607966457,
      "grad_norm": 1.485634207725525,
      "learning_rate": 5.162473794549266e-06,
      "loss": 1.113,
      "num_input_tokens_seen": 256384,
      "step": 395
    },
    {
      "epoch": 0.20964360587002095,
      "grad_norm": 2.76265549659729,
      "learning_rate": 5.227987421383648e-06,
      "loss": 1.3312,
      "num_input_tokens_seen": 259744,
      "step": 400
    },
    {
      "epoch": 0.21226415094339623,
      "grad_norm": 1.936099886894226,
      "learning_rate": 5.2935010482180295e-06,
      "loss": 1.3715,
      "num_input_tokens_seen": 263456,
      "step": 405
    },
    {
      "epoch": 0.2148846960167715,
      "grad_norm": 1.2638863325119019,
      "learning_rate": 5.359014675052411e-06,
      "loss": 1.0528,
      "num_input_tokens_seen": 266496,
      "step": 410
    },
    {
      "epoch": 0.21750524109014674,
      "grad_norm": 1.4862980842590332,
      "learning_rate": 5.424528301886793e-06,
      "loss": 1.1782,
      "num_input_tokens_seen": 269600,
      "step": 415
    },
    {
      "epoch": 0.22012578616352202,
      "grad_norm": 1.4723613262176514,
      "learning_rate": 5.4900419287211744e-06,
      "loss": 1.2919,
      "num_input_tokens_seen": 273152,
      "step": 420
    },
    {
      "epoch": 0.22274633123689727,
      "grad_norm": 1.3303831815719604,
      "learning_rate": 5.555555555555556e-06,
      "loss": 1.5176,
      "num_input_tokens_seen": 276992,
      "step": 425
    },
    {
      "epoch": 0.22536687631027252,
      "grad_norm": 1.4224932193756104,
      "learning_rate": 5.621069182389937e-06,
      "loss": 1.4355,
      "num_input_tokens_seen": 279296,
      "step": 430
    },
    {
      "epoch": 0.2279874213836478,
      "grad_norm": 0.7546237707138062,
      "learning_rate": 5.686582809224319e-06,
      "loss": 0.93,
      "num_input_tokens_seen": 283296,
      "step": 435
    },
    {
      "epoch": 0.23060796645702306,
      "grad_norm": 2.1513969898223877,
      "learning_rate": 5.752096436058701e-06,
      "loss": 1.5187,
      "num_input_tokens_seen": 286016,
      "step": 440
    },
    {
      "epoch": 0.23322851153039834,
      "grad_norm": 1.6297050714492798,
      "learning_rate": 5.817610062893082e-06,
      "loss": 1.0657,
      "num_input_tokens_seen": 289248,
      "step": 445
    },
    {
      "epoch": 0.2358490566037736,
      "grad_norm": 0.7087370157241821,
      "learning_rate": 5.8831236897274636e-06,
      "loss": 1.4625,
      "num_input_tokens_seen": 292160,
      "step": 450
    },
    {
      "epoch": 0.23846960167714884,
      "grad_norm": 1.4907796382904053,
      "learning_rate": 5.948637316561846e-06,
      "loss": 1.1244,
      "num_input_tokens_seen": 295296,
      "step": 455
    },
    {
      "epoch": 0.24109014675052412,
      "grad_norm": 0.8037194013595581,
      "learning_rate": 6.014150943396227e-06,
      "loss": 1.0574,
      "num_input_tokens_seen": 299360,
      "step": 460
    },
    {
      "epoch": 0.24371069182389937,
      "grad_norm": 1.0380439758300781,
      "learning_rate": 6.079664570230608e-06,
      "loss": 0.9278,
      "num_input_tokens_seen": 303072,
      "step": 465
    },
    {
      "epoch": 0.24633123689727462,
      "grad_norm": 0.9433555603027344,
      "learning_rate": 6.14517819706499e-06,
      "loss": 0.898,
      "num_input_tokens_seen": 307616,
      "step": 470
    },
    {
      "epoch": 0.2489517819706499,
      "grad_norm": 0.9906808137893677,
      "learning_rate": 6.210691823899371e-06,
      "loss": 1.3183,
      "num_input_tokens_seen": 310112,
      "step": 475
    },
    {
      "epoch": 0.25157232704402516,
      "grad_norm": 1.419416069984436,
      "learning_rate": 6.2762054507337535e-06,
      "loss": 1.0244,
      "num_input_tokens_seen": 313472,
      "step": 480
    },
    {
      "epoch": 0.25419287211740044,
      "grad_norm": 1.0347328186035156,
      "learning_rate": 6.341719077568134e-06,
      "loss": 1.1941,
      "num_input_tokens_seen": 316128,
      "step": 485
    },
    {
      "epoch": 0.25681341719077566,
      "grad_norm": 1.1312477588653564,
      "learning_rate": 6.407232704402516e-06,
      "loss": 0.9364,
      "num_input_tokens_seen": 319584,
      "step": 490
    },
    {
      "epoch": 0.25943396226415094,
      "grad_norm": 1.0091118812561035,
      "learning_rate": 6.472746331236898e-06,
      "loss": 0.9761,
      "num_input_tokens_seen": 322496,
      "step": 495
    },
    {
      "epoch": 0.2620545073375262,
      "grad_norm": 1.3198871612548828,
      "learning_rate": 6.538259958071278e-06,
      "loss": 1.5193,
      "num_input_tokens_seen": 325120,
      "step": 500
    },
    {
      "epoch": 0.26467505241090145,
      "grad_norm": 1.3471660614013672,
      "learning_rate": 6.60377358490566e-06,
      "loss": 1.255,
      "num_input_tokens_seen": 327840,
      "step": 505
    },
    {
      "epoch": 0.2672955974842767,
      "grad_norm": 1.2837380170822144,
      "learning_rate": 6.669287211740043e-06,
      "loss": 0.7886,
      "num_input_tokens_seen": 331776,
      "step": 510
    },
    {
      "epoch": 0.269916142557652,
      "grad_norm": 2.0342578887939453,
      "learning_rate": 6.734800838574423e-06,
      "loss": 1.0223,
      "num_input_tokens_seen": 336544,
      "step": 515
    },
    {
      "epoch": 0.27253668763102723,
      "grad_norm": 0.7979872822761536,
      "learning_rate": 6.800314465408805e-06,
      "loss": 1.0708,
      "num_input_tokens_seen": 340768,
      "step": 520
    },
    {
      "epoch": 0.2751572327044025,
      "grad_norm": 1.5256271362304688,
      "learning_rate": 6.865828092243188e-06,
      "loss": 1.3615,
      "num_input_tokens_seen": 343296,
      "step": 525
    },
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 0.9718963503837585,
      "learning_rate": 6.931341719077568e-06,
      "loss": 0.9297,
      "num_input_tokens_seen": 346848,
      "step": 530
    },
    {
      "epoch": 0.280398322851153,
      "grad_norm": 0.9056831002235413,
      "learning_rate": 6.9968553459119495e-06,
      "loss": 1.2102,
      "num_input_tokens_seen": 349632,
      "step": 535
    },
    {
      "epoch": 0.2830188679245283,
      "grad_norm": 0.9807378649711609,
      "learning_rate": 7.062368972746332e-06,
      "loss": 0.941,
      "num_input_tokens_seen": 353024,
      "step": 540
    },
    {
      "epoch": 0.2856394129979036,
      "grad_norm": 1.1657500267028809,
      "learning_rate": 7.127882599580712e-06,
      "loss": 1.2362,
      "num_input_tokens_seen": 355552,
      "step": 545
    },
    {
      "epoch": 0.2882599580712788,
      "grad_norm": 1.4837124347686768,
      "learning_rate": 7.1933962264150945e-06,
      "loss": 0.8085,
      "num_input_tokens_seen": 360128,
      "step": 550
    },
    {
      "epoch": 0.2908805031446541,
      "grad_norm": 0.7896787524223328,
      "learning_rate": 7.258909853249477e-06,
      "loss": 0.9297,
      "num_input_tokens_seen": 363360,
      "step": 555
    },
    {
      "epoch": 0.29350104821802936,
      "grad_norm": 1.3429021835327148,
      "learning_rate": 7.324423480083858e-06,
      "loss": 1.1519,
      "num_input_tokens_seen": 366144,
      "step": 560
    },
    {
      "epoch": 0.29612159329140464,
      "grad_norm": 1.4909648895263672,
      "learning_rate": 7.3899371069182395e-06,
      "loss": 0.9932,
      "num_input_tokens_seen": 369216,
      "step": 565
    },
    {
      "epoch": 0.29874213836477986,
      "grad_norm": 1.9894715547561646,
      "learning_rate": 7.455450733752621e-06,
      "loss": 1.008,
      "num_input_tokens_seen": 372512,
      "step": 570
    },
    {
      "epoch": 0.30136268343815514,
      "grad_norm": 1.2826766967773438,
      "learning_rate": 7.520964360587003e-06,
      "loss": 1.2519,
      "num_input_tokens_seen": 374976,
      "step": 575
    },
    {
      "epoch": 0.3039832285115304,
      "grad_norm": 2.847470283508301,
      "learning_rate": 7.586477987421384e-06,
      "loss": 1.0918,
      "num_input_tokens_seen": 377632,
      "step": 580
    },
    {
      "epoch": 0.30660377358490565,
      "grad_norm": 0.8619402647018433,
      "learning_rate": 7.651991614255765e-06,
      "loss": 1.1725,
      "num_input_tokens_seen": 380768,
      "step": 585
    },
    {
      "epoch": 0.30922431865828093,
      "grad_norm": 1.0944697856903076,
      "learning_rate": 7.717505241090148e-06,
      "loss": 1.0223,
      "num_input_tokens_seen": 383520,
      "step": 590
    },
    {
      "epoch": 0.3118448637316562,
      "grad_norm": 3.131998300552368,
      "learning_rate": 7.783018867924528e-06,
      "loss": 1.2908,
      "num_input_tokens_seen": 386496,
      "step": 595
    },
    {
      "epoch": 0.31446540880503143,
      "grad_norm": 1.3699122667312622,
      "learning_rate": 7.848532494758911e-06,
      "loss": 0.982,
      "num_input_tokens_seen": 390464,
      "step": 600
    },
    {
      "epoch": 0.3170859538784067,
      "grad_norm": 1.3089417219161987,
      "learning_rate": 7.914046121593292e-06,
      "loss": 0.8656,
      "num_input_tokens_seen": 393440,
      "step": 605
    },
    {
      "epoch": 0.319706498951782,
      "grad_norm": 0.8193648457527161,
      "learning_rate": 7.979559748427674e-06,
      "loss": 0.8299,
      "num_input_tokens_seen": 398336,
      "step": 610
    },
    {
      "epoch": 0.3223270440251572,
      "grad_norm": 1.7752243280410767,
      "learning_rate": 8.045073375262055e-06,
      "loss": 0.9578,
      "num_input_tokens_seen": 401504,
      "step": 615
    },
    {
      "epoch": 0.3249475890985325,
      "grad_norm": 1.8802986145019531,
      "learning_rate": 8.110587002096436e-06,
      "loss": 0.9046,
      "num_input_tokens_seen": 404864,
      "step": 620
    },
    {
      "epoch": 0.3275681341719078,
      "grad_norm": 0.6981635093688965,
      "learning_rate": 8.176100628930818e-06,
      "loss": 0.9816,
      "num_input_tokens_seen": 409056,
      "step": 625
    },
    {
      "epoch": 0.330188679245283,
      "grad_norm": 1.4068737030029297,
      "learning_rate": 8.241614255765199e-06,
      "loss": 0.974,
      "num_input_tokens_seen": 412544,
      "step": 630
    },
    {
      "epoch": 0.3328092243186583,
      "grad_norm": 1.3288977146148682,
      "learning_rate": 8.307127882599582e-06,
      "loss": 0.9609,
      "num_input_tokens_seen": 415520,
      "step": 635
    },
    {
      "epoch": 0.33542976939203356,
      "grad_norm": 1.2594510316848755,
      "learning_rate": 8.372641509433962e-06,
      "loss": 1.6847,
      "num_input_tokens_seen": 417760,
      "step": 640
    },
    {
      "epoch": 0.3380503144654088,
      "grad_norm": 0.768955647945404,
      "learning_rate": 8.438155136268345e-06,
      "loss": 0.9012,
      "num_input_tokens_seen": 421728,
      "step": 645
    },
    {
      "epoch": 0.34067085953878407,
      "grad_norm": 4.102447032928467,
      "learning_rate": 8.503668763102726e-06,
      "loss": 1.0753,
      "num_input_tokens_seen": 425664,
      "step": 650
    },
    {
      "epoch": 0.34329140461215935,
      "grad_norm": 1.5640612840652466,
      "learning_rate": 8.569182389937106e-06,
      "loss": 0.8507,
      "num_input_tokens_seen": 428896,
      "step": 655
    },
    {
      "epoch": 0.34591194968553457,
      "grad_norm": 1.4761862754821777,
      "learning_rate": 8.634696016771489e-06,
      "loss": 1.0146,
      "num_input_tokens_seen": 431744,
      "step": 660
    },
    {
      "epoch": 0.34853249475890985,
      "grad_norm": 1.0080009698867798,
      "learning_rate": 8.70020964360587e-06,
      "loss": 0.8793,
      "num_input_tokens_seen": 435520,
      "step": 665
    },
    {
      "epoch": 0.35115303983228513,
      "grad_norm": 0.932374894618988,
      "learning_rate": 8.765723270440252e-06,
      "loss": 0.8676,
      "num_input_tokens_seen": 438592,
      "step": 670
    },
    {
      "epoch": 0.35377358490566035,
      "grad_norm": 1.0217205286026,
      "learning_rate": 8.831236897274633e-06,
      "loss": 0.8622,
      "num_input_tokens_seen": 441664,
      "step": 675
    },
    {
      "epoch": 0.35639412997903563,
      "grad_norm": 1.900795578956604,
      "learning_rate": 8.896750524109015e-06,
      "loss": 1.0794,
      "num_input_tokens_seen": 444832,
      "step": 680
    },
    {
      "epoch": 0.3590146750524109,
      "grad_norm": 1.0785943269729614,
      "learning_rate": 8.962264150943396e-06,
      "loss": 1.2432,
      "num_input_tokens_seen": 447392,
      "step": 685
    },
    {
      "epoch": 0.36163522012578614,
      "grad_norm": 1.890355110168457,
      "learning_rate": 9.027777777777777e-06,
      "loss": 0.9732,
      "num_input_tokens_seen": 450528,
      "step": 690
    },
    {
      "epoch": 0.3642557651991614,
      "grad_norm": 1.205201268196106,
      "learning_rate": 9.09329140461216e-06,
      "loss": 0.8436,
      "num_input_tokens_seen": 453504,
      "step": 695
    },
    {
      "epoch": 0.3668763102725367,
      "grad_norm": 1.9193824529647827,
      "learning_rate": 9.15880503144654e-06,
      "loss": 1.8513,
      "num_input_tokens_seen": 456288,
      "step": 700
    },
    {
      "epoch": 0.3694968553459119,
      "grad_norm": 1.4613333940505981,
      "learning_rate": 9.224318658280923e-06,
      "loss": 0.7251,
      "num_input_tokens_seen": 459840,
      "step": 705
    },
    {
      "epoch": 0.3721174004192872,
      "grad_norm": 0.7621285319328308,
      "learning_rate": 9.289832285115305e-06,
      "loss": 0.8733,
      "num_input_tokens_seen": 463168,
      "step": 710
    },
    {
      "epoch": 0.3747379454926625,
      "grad_norm": 1.410707950592041,
      "learning_rate": 9.355345911949686e-06,
      "loss": 0.9111,
      "num_input_tokens_seen": 466112,
      "step": 715
    },
    {
      "epoch": 0.37735849056603776,
      "grad_norm": 1.6942334175109863,
      "learning_rate": 9.420859538784067e-06,
      "loss": 0.9724,
      "num_input_tokens_seen": 469568,
      "step": 720
    },
    {
      "epoch": 0.379979035639413,
      "grad_norm": 0.8299980759620667,
      "learning_rate": 9.486373165618449e-06,
      "loss": 0.7596,
      "num_input_tokens_seen": 472896,
      "step": 725
    },
    {
      "epoch": 0.38259958071278827,
      "grad_norm": 1.7489079236984253,
      "learning_rate": 9.551886792452832e-06,
      "loss": 0.8202,
      "num_input_tokens_seen": 475744,
      "step": 730
    },
    {
      "epoch": 0.38522012578616355,
      "grad_norm": 0.4295312166213989,
      "learning_rate": 9.617400419287211e-06,
      "loss": 0.8974,
      "num_input_tokens_seen": 481216,
      "step": 735
    },
    {
      "epoch": 0.38784067085953877,
      "grad_norm": 1.0471233129501343,
      "learning_rate": 9.682914046121595e-06,
      "loss": 0.9847,
      "num_input_tokens_seen": 483488,
      "step": 740
    },
    {
      "epoch": 0.39046121593291405,
      "grad_norm": 1.2192161083221436,
      "learning_rate": 9.748427672955976e-06,
      "loss": 0.9643,
      "num_input_tokens_seen": 487072,
      "step": 745
    },
    {
      "epoch": 0.39308176100628933,
      "grad_norm": 1.0172492265701294,
      "learning_rate": 9.813941299790357e-06,
      "loss": 0.8574,
      "num_input_tokens_seen": 490496,
      "step": 750
    },
    {
      "epoch": 0.39570230607966456,
      "grad_norm": 0.9634575247764587,
      "learning_rate": 9.879454926624739e-06,
      "loss": 0.8111,
      "num_input_tokens_seen": 493824,
      "step": 755
    },
    {
      "epoch": 0.39832285115303984,
      "grad_norm": 1.442833662033081,
      "learning_rate": 9.94496855345912e-06,
      "loss": 1.1684,
      "num_input_tokens_seen": 496192,
      "step": 760
    },
    {
      "epoch": 0.4009433962264151,
      "grad_norm": 0.9089356660842896,
      "learning_rate": 1.0010482180293501e-05,
      "loss": 1.0089,
      "num_input_tokens_seen": 498880,
      "step": 765
    },
    {
      "epoch": 0.40356394129979034,
      "grad_norm": 1.023152232170105,
      "learning_rate": 1.0075995807127883e-05,
      "loss": 0.8748,
      "num_input_tokens_seen": 502016,
      "step": 770
    },
    {
      "epoch": 0.4061844863731656,
      "grad_norm": 0.5461428165435791,
      "learning_rate": 1.0141509433962266e-05,
      "loss": 0.8919,
      "num_input_tokens_seen": 505120,
      "step": 775
    },
    {
      "epoch": 0.4088050314465409,
      "grad_norm": 1.0339337587356567,
      "learning_rate": 1.0207023060796646e-05,
      "loss": 0.8436,
      "num_input_tokens_seen": 508256,
      "step": 780
    },
    {
      "epoch": 0.4114255765199161,
      "grad_norm": 0.9674530625343323,
      "learning_rate": 1.0272536687631027e-05,
      "loss": 0.858,
      "num_input_tokens_seen": 511776,
      "step": 785
    },
    {
      "epoch": 0.4140461215932914,
      "grad_norm": 1.2479768991470337,
      "learning_rate": 1.033805031446541e-05,
      "loss": 1.0142,
      "num_input_tokens_seen": 514464,
      "step": 790
    },
    {
      "epoch": 0.4166666666666667,
      "grad_norm": 0.9496062994003296,
      "learning_rate": 1.040356394129979e-05,
      "loss": 0.8255,
      "num_input_tokens_seen": 518144,
      "step": 795
    },
    {
      "epoch": 0.4192872117400419,
      "grad_norm": 1.463928461074829,
      "learning_rate": 1.0469077568134173e-05,
      "loss": 1.0494,
      "num_input_tokens_seen": 520768,
      "step": 800
    },
    {
      "epoch": 0.4219077568134172,
      "grad_norm": 1.0034734010696411,
      "learning_rate": 1.0534591194968554e-05,
      "loss": 1.008,
      "num_input_tokens_seen": 523808,
      "step": 805
    },
    {
      "epoch": 0.42452830188679247,
      "grad_norm": 1.3796404600143433,
      "learning_rate": 1.0600104821802936e-05,
      "loss": 0.6437,
      "num_input_tokens_seen": 527136,
      "step": 810
    },
    {
      "epoch": 0.4271488469601677,
      "grad_norm": 1.3821794986724854,
      "learning_rate": 1.0665618448637317e-05,
      "loss": 0.9278,
      "num_input_tokens_seen": 530240,
      "step": 815
    },
    {
      "epoch": 0.429769392033543,
      "grad_norm": 1.8845293521881104,
      "learning_rate": 1.0731132075471698e-05,
      "loss": 0.8469,
      "num_input_tokens_seen": 534496,
      "step": 820
    },
    {
      "epoch": 0.43238993710691825,
      "grad_norm": 2.3409781455993652,
      "learning_rate": 1.079664570230608e-05,
      "loss": 0.8628,
      "num_input_tokens_seen": 538624,
      "step": 825
    },
    {
      "epoch": 0.4350104821802935,
      "grad_norm": 1.4771867990493774,
      "learning_rate": 1.0862159329140461e-05,
      "loss": 0.8884,
      "num_input_tokens_seen": 542112,
      "step": 830
    },
    {
      "epoch": 0.43763102725366876,
      "grad_norm": 1.0132925510406494,
      "learning_rate": 1.0927672955974844e-05,
      "loss": 0.9381,
      "num_input_tokens_seen": 545408,
      "step": 835
    },
    {
      "epoch": 0.44025157232704404,
      "grad_norm": 0.5600353479385376,
      "learning_rate": 1.0993186582809224e-05,
      "loss": 0.7983,
      "num_input_tokens_seen": 548416,
      "step": 840
    },
    {
      "epoch": 0.44287211740041926,
      "grad_norm": 1.1033308506011963,
      "learning_rate": 1.1058700209643607e-05,
      "loss": 0.8325,
      "num_input_tokens_seen": 551264,
      "step": 845
    },
    {
      "epoch": 0.44549266247379454,
      "grad_norm": 0.6903501749038696,
      "learning_rate": 1.1124213836477988e-05,
      "loss": 0.6804,
      "num_input_tokens_seen": 554592,
      "step": 850
    },
    {
      "epoch": 0.4481132075471698,
      "grad_norm": 1.560505747795105,
      "learning_rate": 1.118972746331237e-05,
      "loss": 0.6151,
      "num_input_tokens_seen": 558304,
      "step": 855
    },
    {
      "epoch": 0.45073375262054505,
      "grad_norm": 1.060202956199646,
      "learning_rate": 1.1255241090146751e-05,
      "loss": 0.8777,
      "num_input_tokens_seen": 561248,
      "step": 860
    },
    {
      "epoch": 0.4533542976939203,
      "grad_norm": 0.7758122682571411,
      "learning_rate": 1.1320754716981132e-05,
      "loss": 0.6354,
      "num_input_tokens_seen": 564352,
      "step": 865
    },
    {
      "epoch": 0.4559748427672956,
      "grad_norm": 1.84959077835083,
      "learning_rate": 1.1386268343815514e-05,
      "loss": 0.7525,
      "num_input_tokens_seen": 567712,
      "step": 870
    },
    {
      "epoch": 0.4585953878406709,
      "grad_norm": 1.3714218139648438,
      "learning_rate": 1.1451781970649895e-05,
      "loss": 0.8766,
      "num_input_tokens_seen": 570240,
      "step": 875
    },
    {
      "epoch": 0.4612159329140461,
      "grad_norm": 1.809556245803833,
      "learning_rate": 1.1517295597484278e-05,
      "loss": 0.9219,
      "num_input_tokens_seen": 572768,
      "step": 880
    },
    {
      "epoch": 0.4638364779874214,
      "grad_norm": 0.8189603686332703,
      "learning_rate": 1.158280922431866e-05,
      "loss": 0.7997,
      "num_input_tokens_seen": 576512,
      "step": 885
    },
    {
      "epoch": 0.46645702306079667,
      "grad_norm": 1.0600937604904175,
      "learning_rate": 1.164832285115304e-05,
      "loss": 0.8581,
      "num_input_tokens_seen": 579264,
      "step": 890
    },
    {
      "epoch": 0.4690775681341719,
      "grad_norm": 0.7340589165687561,
      "learning_rate": 1.1713836477987422e-05,
      "loss": 0.9608,
      "num_input_tokens_seen": 581440,
      "step": 895
    },
    {
      "epoch": 0.4716981132075472,
      "grad_norm": 0.6603421568870544,
      "learning_rate": 1.1779350104821804e-05,
      "loss": 0.5699,
      "num_input_tokens_seen": 584544,
      "step": 900
    },
    {
      "epoch": 0.47431865828092246,
      "grad_norm": 1.0413669347763062,
      "learning_rate": 1.1844863731656185e-05,
      "loss": 0.787,
      "num_input_tokens_seen": 588160,
      "step": 905
    },
    {
      "epoch": 0.4769392033542977,
      "grad_norm": 1.9491227865219116,
      "learning_rate": 1.1910377358490566e-05,
      "loss": 0.8652,
      "num_input_tokens_seen": 590400,
      "step": 910
    },
    {
      "epoch": 0.47955974842767296,
      "grad_norm": 0.9039000868797302,
      "learning_rate": 1.197589098532495e-05,
      "loss": 0.6771,
      "num_input_tokens_seen": 593440,
      "step": 915
    },
    {
      "epoch": 0.48218029350104824,
      "grad_norm": 1.1810554265975952,
      "learning_rate": 1.204140461215933e-05,
      "loss": 0.7917,
      "num_input_tokens_seen": 596480,
      "step": 920
    },
    {
      "epoch": 0.48480083857442346,
      "grad_norm": 0.9129461646080017,
      "learning_rate": 1.210691823899371e-05,
      "loss": 0.7332,
      "num_input_tokens_seen": 599392,
      "step": 925
    },
    {
      "epoch": 0.48742138364779874,
      "grad_norm": 0.8661773800849915,
      "learning_rate": 1.2172431865828094e-05,
      "loss": 0.6215,
      "num_input_tokens_seen": 602432,
      "step": 930
    },
    {
      "epoch": 0.490041928721174,
      "grad_norm": 1.2933578491210938,
      "learning_rate": 1.2237945492662473e-05,
      "loss": 0.8049,
      "num_input_tokens_seen": 604512,
      "step": 935
    },
    {
      "epoch": 0.49266247379454925,
      "grad_norm": 0.5134191513061523,
      "learning_rate": 1.2303459119496856e-05,
      "loss": 0.6622,
      "num_input_tokens_seen": 608864,
      "step": 940
    },
    {
      "epoch": 0.49528301886792453,
      "grad_norm": 1.039077877998352,
      "learning_rate": 1.2368972746331238e-05,
      "loss": 0.6706,
      "num_input_tokens_seen": 612448,
      "step": 945
    },
    {
      "epoch": 0.4979035639412998,
      "grad_norm": 4.265715599060059,
      "learning_rate": 1.243448637316562e-05,
      "loss": 0.9209,
      "num_input_tokens_seen": 614944,
      "step": 950
    },
    {
      "epoch": 0.500524109014675,
      "grad_norm": 1.0462404489517212,
      "learning_rate": 1.25e-05,
      "loss": 0.7243,
      "num_input_tokens_seen": 617472,
      "step": 955
    },
    {
      "epoch": 0.5031446540880503,
      "grad_norm": 1.227678894996643,
      "learning_rate": 1.2565513626834382e-05,
      "loss": 0.668,
      "num_input_tokens_seen": 620192,
      "step": 960
    },
    {
      "epoch": 0.5057651991614256,
      "grad_norm": 0.6380628347396851,
      "learning_rate": 1.2631027253668765e-05,
      "loss": 0.738,
      "num_input_tokens_seen": 623232,
      "step": 965
    },
    {
      "epoch": 0.5083857442348009,
      "grad_norm": 1.025099754333496,
      "learning_rate": 1.2696540880503146e-05,
      "loss": 0.8766,
      "num_input_tokens_seen": 626016,
      "step": 970
    },
    {
      "epoch": 0.5110062893081762,
      "grad_norm": 0.7771926522254944,
      "learning_rate": 1.2762054507337526e-05,
      "loss": 0.658,
      "num_input_tokens_seen": 629472,
      "step": 975
    },
    {
      "epoch": 0.5136268343815513,
      "grad_norm": 1.1121550798416138,
      "learning_rate": 1.2827568134171907e-05,
      "loss": 0.8124,
      "num_input_tokens_seen": 632608,
      "step": 980
    },
    {
      "epoch": 0.5162473794549266,
      "grad_norm": 1.082263469696045,
      "learning_rate": 1.289308176100629e-05,
      "loss": 0.6304,
      "num_input_tokens_seen": 636192,
      "step": 985
    },
    {
      "epoch": 0.5188679245283019,
      "grad_norm": 0.6882573962211609,
      "learning_rate": 1.2958595387840672e-05,
      "loss": 0.6376,
      "num_input_tokens_seen": 640544,
      "step": 990
    },
    {
      "epoch": 0.5214884696016772,
      "grad_norm": 0.9138583540916443,
      "learning_rate": 1.3024109014675053e-05,
      "loss": 0.6243,
      "num_input_tokens_seen": 644448,
      "step": 995
    },
    {
      "epoch": 0.5241090146750524,
      "grad_norm": 0.6401057243347168,
      "learning_rate": 1.3089622641509436e-05,
      "loss": 0.6284,
      "num_input_tokens_seen": 648288,
      "step": 1000
    },
    {
      "epoch": 0.5267295597484277,
      "grad_norm": 1.4025288820266724,
      "learning_rate": 1.3155136268343814e-05,
      "loss": 0.8307,
      "num_input_tokens_seen": 651296,
      "step": 1005
    },
    {
      "epoch": 0.5293501048218029,
      "grad_norm": 0.5005475282669067,
      "learning_rate": 1.3220649895178197e-05,
      "loss": 0.6545,
      "num_input_tokens_seen": 654176,
      "step": 1010
    },
    {
      "epoch": 0.5319706498951782,
      "grad_norm": 0.6839974522590637,
      "learning_rate": 1.3286163522012579e-05,
      "loss": 0.6775,
      "num_input_tokens_seen": 657664,
      "step": 1015
    },
    {
      "epoch": 0.5345911949685535,
      "grad_norm": 1.199899435043335,
      "learning_rate": 1.3351677148846962e-05,
      "loss": 0.8223,
      "num_input_tokens_seen": 660064,
      "step": 1020
    },
    {
      "epoch": 0.5372117400419287,
      "grad_norm": 1.1473360061645508,
      "learning_rate": 1.3417190775681343e-05,
      "loss": 0.7799,
      "num_input_tokens_seen": 662880,
      "step": 1025
    },
    {
      "epoch": 0.539832285115304,
      "grad_norm": 0.7052485346794128,
      "learning_rate": 1.3482704402515725e-05,
      "loss": 0.6218,
      "num_input_tokens_seen": 665920,
      "step": 1030
    },
    {
      "epoch": 0.5424528301886793,
      "grad_norm": 1.0897216796875,
      "learning_rate": 1.3548218029350104e-05,
      "loss": 0.6189,
      "num_input_tokens_seen": 668864,
      "step": 1035
    },
    {
      "epoch": 0.5450733752620545,
      "grad_norm": 0.8728546500205994,
      "learning_rate": 1.3613731656184486e-05,
      "loss": 0.5946,
      "num_input_tokens_seen": 671776,
      "step": 1040
    },
    {
      "epoch": 0.5476939203354297,
      "grad_norm": 0.8971770405769348,
      "learning_rate": 1.3679245283018869e-05,
      "loss": 0.6877,
      "num_input_tokens_seen": 674912,
      "step": 1045
    },
    {
      "epoch": 0.550314465408805,
      "grad_norm": 0.8158504962921143,
      "learning_rate": 1.374475890985325e-05,
      "loss": 0.5254,
      "num_input_tokens_seen": 678368,
      "step": 1050
    },
    {
      "epoch": 0.5529350104821803,
      "grad_norm": 0.4873327612876892,
      "learning_rate": 1.3810272536687632e-05,
      "loss": 0.7425,
      "num_input_tokens_seen": 682336,
      "step": 1055
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 1.6114007234573364,
      "learning_rate": 1.3875786163522015e-05,
      "loss": 0.641,
      "num_input_tokens_seen": 685760,
      "step": 1060
    },
    {
      "epoch": 0.5581761006289309,
      "grad_norm": 1.3626179695129395,
      "learning_rate": 1.3941299790356394e-05,
      "loss": 0.5746,
      "num_input_tokens_seen": 688480,
      "step": 1065
    },
    {
      "epoch": 0.560796645702306,
      "grad_norm": 0.8061174154281616,
      "learning_rate": 1.4006813417190776e-05,
      "loss": 0.79,
      "num_input_tokens_seen": 692064,
      "step": 1070
    },
    {
      "epoch": 0.5634171907756813,
      "grad_norm": 1.5884085893630981,
      "learning_rate": 1.4072327044025157e-05,
      "loss": 0.7777,
      "num_input_tokens_seen": 694464,
      "step": 1075
    },
    {
      "epoch": 0.5660377358490566,
      "grad_norm": 1.0193461179733276,
      "learning_rate": 1.413784067085954e-05,
      "loss": 0.6883,
      "num_input_tokens_seen": 697344,
      "step": 1080
    },
    {
      "epoch": 0.5686582809224319,
      "grad_norm": 1.120331883430481,
      "learning_rate": 1.4203354297693922e-05,
      "loss": 0.6932,
      "num_input_tokens_seen": 700256,
      "step": 1085
    },
    {
      "epoch": 0.5712788259958071,
      "grad_norm": 0.7114775776863098,
      "learning_rate": 1.4268867924528303e-05,
      "loss": 0.6559,
      "num_input_tokens_seen": 703968,
      "step": 1090
    },
    {
      "epoch": 0.5738993710691824,
      "grad_norm": 0.6708186864852905,
      "learning_rate": 1.4334381551362683e-05,
      "loss": 0.6365,
      "num_input_tokens_seen": 708160,
      "step": 1095
    },
    {
      "epoch": 0.5765199161425576,
      "grad_norm": 0.7196803689002991,
      "learning_rate": 1.4399895178197064e-05,
      "loss": 0.873,
      "num_input_tokens_seen": 710976,
      "step": 1100
    },
    {
      "epoch": 0.5791404612159329,
      "grad_norm": 2.156536817550659,
      "learning_rate": 1.4465408805031447e-05,
      "loss": 0.8462,
      "num_input_tokens_seen": 713696,
      "step": 1105
    },
    {
      "epoch": 0.5817610062893082,
      "grad_norm": 1.222192406654358,
      "learning_rate": 1.4530922431865828e-05,
      "loss": 0.6461,
      "num_input_tokens_seen": 716608,
      "step": 1110
    },
    {
      "epoch": 0.5843815513626834,
      "grad_norm": 0.7029736638069153,
      "learning_rate": 1.4596436058700211e-05,
      "loss": 0.6218,
      "num_input_tokens_seen": 719488,
      "step": 1115
    },
    {
      "epoch": 0.5870020964360587,
      "grad_norm": 0.6068930625915527,
      "learning_rate": 1.4661949685534593e-05,
      "loss": 0.549,
      "num_input_tokens_seen": 722304,
      "step": 1120
    },
    {
      "epoch": 0.589622641509434,
      "grad_norm": 1.0509752035140991,
      "learning_rate": 1.4727463312368974e-05,
      "loss": 0.5871,
      "num_input_tokens_seen": 725472,
      "step": 1125
    },
    {
      "epoch": 0.5922431865828093,
      "grad_norm": 0.7260472774505615,
      "learning_rate": 1.4792976939203354e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 728896,
      "step": 1130
    },
    {
      "epoch": 0.5948637316561844,
      "grad_norm": 0.5641947984695435,
      "learning_rate": 1.4858490566037735e-05,
      "loss": 0.5847,
      "num_input_tokens_seen": 732000,
      "step": 1135
    },
    {
      "epoch": 0.5974842767295597,
      "grad_norm": 1.0996497869491577,
      "learning_rate": 1.4924004192872118e-05,
      "loss": 0.6957,
      "num_input_tokens_seen": 735328,
      "step": 1140
    },
    {
      "epoch": 0.600104821802935,
      "grad_norm": 0.7035921216011047,
      "learning_rate": 1.49895178197065e-05,
      "loss": 0.5218,
      "num_input_tokens_seen": 738528,
      "step": 1145
    },
    {
      "epoch": 0.6027253668763103,
      "grad_norm": 0.6831376552581787,
      "learning_rate": 1.5055031446540883e-05,
      "loss": 0.7888,
      "num_input_tokens_seen": 741376,
      "step": 1150
    },
    {
      "epoch": 0.6053459119496856,
      "grad_norm": 0.8473286628723145,
      "learning_rate": 1.5120545073375264e-05,
      "loss": 0.6678,
      "num_input_tokens_seen": 743872,
      "step": 1155
    },
    {
      "epoch": 0.6079664570230608,
      "grad_norm": 0.8363293409347534,
      "learning_rate": 1.5186058700209644e-05,
      "loss": 0.7054,
      "num_input_tokens_seen": 746720,
      "step": 1160
    },
    {
      "epoch": 0.610587002096436,
      "grad_norm": 0.794966995716095,
      "learning_rate": 1.5251572327044025e-05,
      "loss": 0.7134,
      "num_input_tokens_seen": 750240,
      "step": 1165
    },
    {
      "epoch": 0.6132075471698113,
      "grad_norm": 0.7972670197486877,
      "learning_rate": 1.5317085953878407e-05,
      "loss": 0.6216,
      "num_input_tokens_seen": 753344,
      "step": 1170
    },
    {
      "epoch": 0.6158280922431866,
      "grad_norm": 0.7607865929603577,
      "learning_rate": 1.538259958071279e-05,
      "loss": 0.7061,
      "num_input_tokens_seen": 755968,
      "step": 1175
    },
    {
      "epoch": 0.6184486373165619,
      "grad_norm": 0.44406658411026,
      "learning_rate": 1.5448113207547173e-05,
      "loss": 0.6997,
      "num_input_tokens_seen": 759520,
      "step": 1180
    },
    {
      "epoch": 0.6210691823899371,
      "grad_norm": 0.7881073355674744,
      "learning_rate": 1.5513626834381552e-05,
      "loss": 0.6185,
      "num_input_tokens_seen": 762048,
      "step": 1185
    },
    {
      "epoch": 0.6236897274633124,
      "grad_norm": 0.9658064842224121,
      "learning_rate": 1.5579140461215932e-05,
      "loss": 0.8012,
      "num_input_tokens_seen": 765504,
      "step": 1190
    },
    {
      "epoch": 0.6263102725366876,
      "grad_norm": 0.5292953252792358,
      "learning_rate": 1.5644654088050315e-05,
      "loss": 0.5988,
      "num_input_tokens_seen": 768192,
      "step": 1195
    },
    {
      "epoch": 0.6289308176100629,
      "grad_norm": 0.6328336596488953,
      "learning_rate": 1.5710167714884695e-05,
      "loss": 0.655,
      "num_input_tokens_seen": 771168,
      "step": 1200
    },
    {
      "epoch": 0.6315513626834381,
      "grad_norm": 0.6113349795341492,
      "learning_rate": 1.5775681341719078e-05,
      "loss": 0.386,
      "num_input_tokens_seen": 777504,
      "step": 1205
    },
    {
      "epoch": 0.6341719077568134,
      "grad_norm": 1.0411109924316406,
      "learning_rate": 1.584119496855346e-05,
      "loss": 0.6013,
      "num_input_tokens_seen": 780512,
      "step": 1210
    },
    {
      "epoch": 0.6367924528301887,
      "grad_norm": 0.7733576893806458,
      "learning_rate": 1.5906708595387844e-05,
      "loss": 0.6492,
      "num_input_tokens_seen": 784192,
      "step": 1215
    },
    {
      "epoch": 0.639412997903564,
      "grad_norm": 1.3870757818222046,
      "learning_rate": 1.597222222222222e-05,
      "loss": 0.708,
      "num_input_tokens_seen": 787552,
      "step": 1220
    },
    {
      "epoch": 0.6420335429769392,
      "grad_norm": 0.4692288637161255,
      "learning_rate": 1.6037735849056604e-05,
      "loss": 0.7912,
      "num_input_tokens_seen": 790720,
      "step": 1225
    },
    {
      "epoch": 0.6446540880503144,
      "grad_norm": 1.0507304668426514,
      "learning_rate": 1.6103249475890987e-05,
      "loss": 0.6971,
      "num_input_tokens_seen": 793696,
      "step": 1230
    },
    {
      "epoch": 0.6472746331236897,
      "grad_norm": 0.7973852753639221,
      "learning_rate": 1.6168763102725366e-05,
      "loss": 0.6986,
      "num_input_tokens_seen": 796640,
      "step": 1235
    },
    {
      "epoch": 0.649895178197065,
      "grad_norm": 0.8996561765670776,
      "learning_rate": 1.623427672955975e-05,
      "loss": 0.615,
      "num_input_tokens_seen": 799648,
      "step": 1240
    },
    {
      "epoch": 0.6525157232704403,
      "grad_norm": 0.8491382002830505,
      "learning_rate": 1.6299790356394132e-05,
      "loss": 0.4334,
      "num_input_tokens_seen": 802496,
      "step": 1245
    },
    {
      "epoch": 0.6551362683438156,
      "grad_norm": 0.8419046998023987,
      "learning_rate": 1.6365303983228512e-05,
      "loss": 0.6735,
      "num_input_tokens_seen": 805760,
      "step": 1250
    },
    {
      "epoch": 0.6577568134171907,
      "grad_norm": 1.2534301280975342,
      "learning_rate": 1.6430817610062892e-05,
      "loss": 0.71,
      "num_input_tokens_seen": 808832,
      "step": 1255
    },
    {
      "epoch": 0.660377358490566,
      "grad_norm": 0.8097022175788879,
      "learning_rate": 1.6496331236897275e-05,
      "loss": 0.5617,
      "num_input_tokens_seen": 812256,
      "step": 1260
    },
    {
      "epoch": 0.6629979035639413,
      "grad_norm": 0.8039873838424683,
      "learning_rate": 1.6561844863731658e-05,
      "loss": 0.5755,
      "num_input_tokens_seen": 815680,
      "step": 1265
    },
    {
      "epoch": 0.6656184486373166,
      "grad_norm": 0.5346124172210693,
      "learning_rate": 1.6627358490566038e-05,
      "loss": 0.5148,
      "num_input_tokens_seen": 818400,
      "step": 1270
    },
    {
      "epoch": 0.6682389937106918,
      "grad_norm": 0.6737653613090515,
      "learning_rate": 1.669287211740042e-05,
      "loss": 0.6219,
      "num_input_tokens_seen": 821536,
      "step": 1275
    },
    {
      "epoch": 0.6708595387840671,
      "grad_norm": 0.9566025137901306,
      "learning_rate": 1.6758385744234804e-05,
      "loss": 0.549,
      "num_input_tokens_seen": 825024,
      "step": 1280
    },
    {
      "epoch": 0.6734800838574424,
      "grad_norm": 0.3910340666770935,
      "learning_rate": 1.6823899371069183e-05,
      "loss": 0.5221,
      "num_input_tokens_seen": 829248,
      "step": 1285
    },
    {
      "epoch": 0.6761006289308176,
      "grad_norm": 0.603003740310669,
      "learning_rate": 1.6889412997903563e-05,
      "loss": 0.5741,
      "num_input_tokens_seen": 831872,
      "step": 1290
    },
    {
      "epoch": 0.6787211740041929,
      "grad_norm": 0.6108189821243286,
      "learning_rate": 1.6954926624737946e-05,
      "loss": 0.742,
      "num_input_tokens_seen": 834880,
      "step": 1295
    },
    {
      "epoch": 0.6813417190775681,
      "grad_norm": 0.5574046969413757,
      "learning_rate": 1.702044025157233e-05,
      "loss": 0.7144,
      "num_input_tokens_seen": 837824,
      "step": 1300
    },
    {
      "epoch": 0.6839622641509434,
      "grad_norm": 0.6748602390289307,
      "learning_rate": 1.708595387840671e-05,
      "loss": 0.695,
      "num_input_tokens_seen": 840896,
      "step": 1305
    },
    {
      "epoch": 0.6865828092243187,
      "grad_norm": 0.7146056294441223,
      "learning_rate": 1.7151467505241092e-05,
      "loss": 0.7697,
      "num_input_tokens_seen": 844000,
      "step": 1310
    },
    {
      "epoch": 0.689203354297694,
      "grad_norm": 0.26959550380706787,
      "learning_rate": 1.7216981132075472e-05,
      "loss": 0.5579,
      "num_input_tokens_seen": 847840,
      "step": 1315
    },
    {
      "epoch": 0.6918238993710691,
      "grad_norm": 0.773445725440979,
      "learning_rate": 1.7282494758909855e-05,
      "loss": 0.7351,
      "num_input_tokens_seen": 850976,
      "step": 1320
    },
    {
      "epoch": 0.6944444444444444,
      "grad_norm": 0.6624552607536316,
      "learning_rate": 1.7348008385744234e-05,
      "loss": 0.5582,
      "num_input_tokens_seen": 855616,
      "step": 1325
    },
    {
      "epoch": 0.6970649895178197,
      "grad_norm": 0.5317797660827637,
      "learning_rate": 1.7413522012578618e-05,
      "loss": 0.3995,
      "num_input_tokens_seen": 858528,
      "step": 1330
    },
    {
      "epoch": 0.699685534591195,
      "grad_norm": 0.8354832530021667,
      "learning_rate": 1.7479035639413e-05,
      "loss": 0.6065,
      "num_input_tokens_seen": 862144,
      "step": 1335
    },
    {
      "epoch": 0.7023060796645703,
      "grad_norm": 0.41073641180992126,
      "learning_rate": 1.754454926624738e-05,
      "loss": 0.6612,
      "num_input_tokens_seen": 865440,
      "step": 1340
    },
    {
      "epoch": 0.7049266247379455,
      "grad_norm": 0.4931684136390686,
      "learning_rate": 1.761006289308176e-05,
      "loss": 0.6392,
      "num_input_tokens_seen": 868992,
      "step": 1345
    },
    {
      "epoch": 0.7075471698113207,
      "grad_norm": 1.0645933151245117,
      "learning_rate": 1.7675576519916143e-05,
      "loss": 0.5838,
      "num_input_tokens_seen": 871968,
      "step": 1350
    },
    {
      "epoch": 0.710167714884696,
      "grad_norm": 0.7632243633270264,
      "learning_rate": 1.7741090146750526e-05,
      "loss": 0.5977,
      "num_input_tokens_seen": 874912,
      "step": 1355
    },
    {
      "epoch": 0.7127882599580713,
      "grad_norm": 0.4854307472705841,
      "learning_rate": 1.7806603773584906e-05,
      "loss": 0.6099,
      "num_input_tokens_seen": 878112,
      "step": 1360
    },
    {
      "epoch": 0.7154088050314465,
      "grad_norm": 1.6117717027664185,
      "learning_rate": 1.787211740041929e-05,
      "loss": 0.5992,
      "num_input_tokens_seen": 880928,
      "step": 1365
    },
    {
      "epoch": 0.7180293501048218,
      "grad_norm": 0.5776271224021912,
      "learning_rate": 1.7937631027253672e-05,
      "loss": 0.594,
      "num_input_tokens_seen": 884832,
      "step": 1370
    },
    {
      "epoch": 0.7206498951781971,
      "grad_norm": 0.4581120014190674,
      "learning_rate": 1.8003144654088048e-05,
      "loss": 0.5634,
      "num_input_tokens_seen": 888800,
      "step": 1375
    },
    {
      "epoch": 0.7232704402515723,
      "grad_norm": 0.5944256782531738,
      "learning_rate": 1.806865828092243e-05,
      "loss": 0.587,
      "num_input_tokens_seen": 892416,
      "step": 1380
    },
    {
      "epoch": 0.7258909853249476,
      "grad_norm": 0.9223570823669434,
      "learning_rate": 1.8134171907756814e-05,
      "loss": 0.6955,
      "num_input_tokens_seen": 895840,
      "step": 1385
    },
    {
      "epoch": 0.7285115303983228,
      "grad_norm": 0.5619350075721741,
      "learning_rate": 1.8199685534591197e-05,
      "loss": 0.6141,
      "num_input_tokens_seen": 899392,
      "step": 1390
    },
    {
      "epoch": 0.7311320754716981,
      "grad_norm": 0.5632548332214355,
      "learning_rate": 1.8265199161425577e-05,
      "loss": 0.5578,
      "num_input_tokens_seen": 903264,
      "step": 1395
    },
    {
      "epoch": 0.7337526205450734,
      "grad_norm": 0.6138390898704529,
      "learning_rate": 1.833071278825996e-05,
      "loss": 0.6471,
      "num_input_tokens_seen": 906016,
      "step": 1400
    },
    {
      "epoch": 0.7363731656184487,
      "grad_norm": 1.1559712886810303,
      "learning_rate": 1.839622641509434e-05,
      "loss": 0.6917,
      "num_input_tokens_seen": 909152,
      "step": 1405
    },
    {
      "epoch": 0.7389937106918238,
      "grad_norm": 0.5470163226127625,
      "learning_rate": 1.846174004192872e-05,
      "loss": 0.6649,
      "num_input_tokens_seen": 912416,
      "step": 1410
    },
    {
      "epoch": 0.7416142557651991,
      "grad_norm": 0.7190682291984558,
      "learning_rate": 1.8527253668763103e-05,
      "loss": 0.7838,
      "num_input_tokens_seen": 915616,
      "step": 1415
    },
    {
      "epoch": 0.7442348008385744,
      "grad_norm": 1.1374591588974,
      "learning_rate": 1.8592767295597486e-05,
      "loss": 0.6336,
      "num_input_tokens_seen": 921568,
      "step": 1420
    },
    {
      "epoch": 0.7468553459119497,
      "grad_norm": 0.49394553899765015,
      "learning_rate": 1.865828092243187e-05,
      "loss": 0.5353,
      "num_input_tokens_seen": 924864,
      "step": 1425
    },
    {
      "epoch": 0.749475890985325,
      "grad_norm": 1.001574993133545,
      "learning_rate": 1.872379454926625e-05,
      "loss": 0.5813,
      "num_input_tokens_seen": 927968,
      "step": 1430
    },
    {
      "epoch": 0.7520964360587002,
      "grad_norm": 0.7330270409584045,
      "learning_rate": 1.878930817610063e-05,
      "loss": 0.5945,
      "num_input_tokens_seen": 931584,
      "step": 1435
    },
    {
      "epoch": 0.7547169811320755,
      "grad_norm": 0.9106690883636475,
      "learning_rate": 1.885482180293501e-05,
      "loss": 0.585,
      "num_input_tokens_seen": 934816,
      "step": 1440
    },
    {
      "epoch": 0.7573375262054507,
      "grad_norm": 0.8550238013267517,
      "learning_rate": 1.892033542976939e-05,
      "loss": 0.6382,
      "num_input_tokens_seen": 937856,
      "step": 1445
    },
    {
      "epoch": 0.759958071278826,
      "grad_norm": 0.8097718954086304,
      "learning_rate": 1.8985849056603774e-05,
      "loss": 0.62,
      "num_input_tokens_seen": 945632,
      "step": 1450
    },
    {
      "epoch": 0.7625786163522013,
      "grad_norm": 0.7013465166091919,
      "learning_rate": 1.9051362683438157e-05,
      "loss": 0.7001,
      "num_input_tokens_seen": 948608,
      "step": 1455
    },
    {
      "epoch": 0.7651991614255765,
      "grad_norm": 0.9960947632789612,
      "learning_rate": 1.9116876310272537e-05,
      "loss": 0.609,
      "num_input_tokens_seen": 951488,
      "step": 1460
    },
    {
      "epoch": 0.7678197064989518,
      "grad_norm": 0.7014427185058594,
      "learning_rate": 1.918238993710692e-05,
      "loss": 0.631,
      "num_input_tokens_seen": 955072,
      "step": 1465
    },
    {
      "epoch": 0.7704402515723271,
      "grad_norm": 0.9242023229598999,
      "learning_rate": 1.92479035639413e-05,
      "loss": 0.6494,
      "num_input_tokens_seen": 958272,
      "step": 1470
    },
    {
      "epoch": 0.7730607966457023,
      "grad_norm": 0.5057815313339233,
      "learning_rate": 1.9313417190775683e-05,
      "loss": 0.7301,
      "num_input_tokens_seen": 961888,
      "step": 1475
    },
    {
      "epoch": 0.7756813417190775,
      "grad_norm": 1.300520896911621,
      "learning_rate": 1.9378930817610062e-05,
      "loss": 0.8154,
      "num_input_tokens_seen": 964256,
      "step": 1480
    },
    {
      "epoch": 0.7783018867924528,
      "grad_norm": 0.7685168981552124,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 0.5963,
      "num_input_tokens_seen": 968224,
      "step": 1485
    },
    {
      "epoch": 0.7809224318658281,
      "grad_norm": 0.4474777281284332,
      "learning_rate": 1.950995807127883e-05,
      "loss": 0.575,
      "num_input_tokens_seen": 971136,
      "step": 1490
    },
    {
      "epoch": 0.7835429769392034,
      "grad_norm": 0.577025294303894,
      "learning_rate": 1.9575471698113208e-05,
      "loss": 0.593,
      "num_input_tokens_seen": 975936,
      "step": 1495
    },
    {
      "epoch": 0.7861635220125787,
      "grad_norm": 0.7897735238075256,
      "learning_rate": 1.9640985324947588e-05,
      "loss": 0.631,
      "num_input_tokens_seen": 979360,
      "step": 1500
    },
    {
      "epoch": 0.7887840670859538,
      "grad_norm": 0.2887365221977234,
      "learning_rate": 1.970649895178197e-05,
      "loss": 0.5511,
      "num_input_tokens_seen": 983616,
      "step": 1505
    },
    {
      "epoch": 0.7914046121593291,
      "grad_norm": 0.5682291388511658,
      "learning_rate": 1.9772012578616354e-05,
      "loss": 0.5986,
      "num_input_tokens_seen": 986944,
      "step": 1510
    },
    {
      "epoch": 0.7940251572327044,
      "grad_norm": 0.739945113658905,
      "learning_rate": 1.9837526205450734e-05,
      "loss": 0.6223,
      "num_input_tokens_seen": 989888,
      "step": 1515
    },
    {
      "epoch": 0.7966457023060797,
      "grad_norm": 0.6289554238319397,
      "learning_rate": 1.9903039832285117e-05,
      "loss": 0.6239,
      "num_input_tokens_seen": 992768,
      "step": 1520
    },
    {
      "epoch": 0.799266247379455,
      "grad_norm": 0.5376354455947876,
      "learning_rate": 1.99685534591195e-05,
      "loss": 0.5101,
      "num_input_tokens_seen": 996896,
      "step": 1525
    },
    {
      "epoch": 0.8018867924528302,
      "grad_norm": 0.4991818070411682,
      "learning_rate": 2.003406708595388e-05,
      "loss": 0.6323,
      "num_input_tokens_seen": 999840,
      "step": 1530
    },
    {
      "epoch": 0.8045073375262054,
      "grad_norm": 0.46798765659332275,
      "learning_rate": 2.009958071278826e-05,
      "loss": 0.5301,
      "num_input_tokens_seen": 1003680,
      "step": 1535
    },
    {
      "epoch": 0.8071278825995807,
      "grad_norm": 0.6094844341278076,
      "learning_rate": 2.0165094339622642e-05,
      "loss": 0.5482,
      "num_input_tokens_seen": 1006656,
      "step": 1540
    },
    {
      "epoch": 0.809748427672956,
      "grad_norm": 2.215724229812622,
      "learning_rate": 2.0230607966457025e-05,
      "loss": 0.5032,
      "num_input_tokens_seen": 1008768,
      "step": 1545
    },
    {
      "epoch": 0.8123689727463312,
      "grad_norm": 0.4083808660507202,
      "learning_rate": 2.0296121593291405e-05,
      "loss": 0.5808,
      "num_input_tokens_seen": 1011968,
      "step": 1550
    },
    {
      "epoch": 0.8149895178197065,
      "grad_norm": 1.2540957927703857,
      "learning_rate": 2.0361635220125788e-05,
      "loss": 0.6436,
      "num_input_tokens_seen": 1018784,
      "step": 1555
    },
    {
      "epoch": 0.8176100628930818,
      "grad_norm": 0.3795686960220337,
      "learning_rate": 2.0427148846960168e-05,
      "loss": 0.4018,
      "num_input_tokens_seen": 1023520,
      "step": 1560
    },
    {
      "epoch": 0.820230607966457,
      "grad_norm": 0.6682443618774414,
      "learning_rate": 2.049266247379455e-05,
      "loss": 0.5324,
      "num_input_tokens_seen": 1027200,
      "step": 1565
    },
    {
      "epoch": 0.8228511530398323,
      "grad_norm": 0.32934293150901794,
      "learning_rate": 2.055817610062893e-05,
      "loss": 0.8955,
      "num_input_tokens_seen": 1030848,
      "step": 1570
    },
    {
      "epoch": 0.8254716981132075,
      "grad_norm": 0.580069363117218,
      "learning_rate": 2.0623689727463314e-05,
      "loss": 0.6222,
      "num_input_tokens_seen": 1035232,
      "step": 1575
    },
    {
      "epoch": 0.8280922431865828,
      "grad_norm": 0.9983121156692505,
      "learning_rate": 2.0689203354297697e-05,
      "loss": 0.5941,
      "num_input_tokens_seen": 1038176,
      "step": 1580
    },
    {
      "epoch": 0.8307127882599581,
      "grad_norm": 1.216804027557373,
      "learning_rate": 2.0754716981132076e-05,
      "loss": 0.5626,
      "num_input_tokens_seen": 1041216,
      "step": 1585
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 0.601603627204895,
      "learning_rate": 2.0820230607966456e-05,
      "loss": 0.6039,
      "num_input_tokens_seen": 1043840,
      "step": 1590
    },
    {
      "epoch": 0.8359538784067087,
      "grad_norm": 0.374472439289093,
      "learning_rate": 2.088574423480084e-05,
      "loss": 0.4898,
      "num_input_tokens_seen": 1046656,
      "step": 1595
    },
    {
      "epoch": 0.8385744234800838,
      "grad_norm": 0.5888379812240601,
      "learning_rate": 2.0951257861635222e-05,
      "loss": 0.5031,
      "num_input_tokens_seen": 1051616,
      "step": 1600
    },
    {
      "epoch": 0.8411949685534591,
      "grad_norm": 0.8933074474334717,
      "learning_rate": 2.1016771488469602e-05,
      "loss": 0.739,
      "num_input_tokens_seen": 1053824,
      "step": 1605
    },
    {
      "epoch": 0.8438155136268344,
      "grad_norm": 0.8649858832359314,
      "learning_rate": 2.1082285115303985e-05,
      "loss": 0.5799,
      "num_input_tokens_seen": 1056512,
      "step": 1610
    },
    {
      "epoch": 0.8464360587002097,
      "grad_norm": 0.34400251507759094,
      "learning_rate": 2.1147798742138368e-05,
      "loss": 0.5544,
      "num_input_tokens_seen": 1060160,
      "step": 1615
    },
    {
      "epoch": 0.8490566037735849,
      "grad_norm": 0.8508387207984924,
      "learning_rate": 2.1213312368972748e-05,
      "loss": 0.738,
      "num_input_tokens_seen": 1062880,
      "step": 1620
    },
    {
      "epoch": 0.8516771488469602,
      "grad_norm": 0.5709040760993958,
      "learning_rate": 2.1278825995807127e-05,
      "loss": 0.6779,
      "num_input_tokens_seen": 1066528,
      "step": 1625
    },
    {
      "epoch": 0.8542976939203354,
      "grad_norm": 0.6534771919250488,
      "learning_rate": 2.134433962264151e-05,
      "loss": 0.4861,
      "num_input_tokens_seen": 1069472,
      "step": 1630
    },
    {
      "epoch": 0.8569182389937107,
      "grad_norm": 0.614721417427063,
      "learning_rate": 2.1409853249475894e-05,
      "loss": 0.5739,
      "num_input_tokens_seen": 1074304,
      "step": 1635
    },
    {
      "epoch": 0.859538784067086,
      "grad_norm": 1.1143416166305542,
      "learning_rate": 2.1475366876310273e-05,
      "loss": 0.539,
      "num_input_tokens_seen": 1077376,
      "step": 1640
    },
    {
      "epoch": 0.8621593291404612,
      "grad_norm": 0.5396094918251038,
      "learning_rate": 2.1540880503144656e-05,
      "loss": 0.5773,
      "num_input_tokens_seen": 1080672,
      "step": 1645
    },
    {
      "epoch": 0.8647798742138365,
      "grad_norm": 0.6036821603775024,
      "learning_rate": 2.160639412997904e-05,
      "loss": 0.6302,
      "num_input_tokens_seen": 1084224,
      "step": 1650
    },
    {
      "epoch": 0.8674004192872118,
      "grad_norm": 0.6878481507301331,
      "learning_rate": 2.1671907756813416e-05,
      "loss": 0.7895,
      "num_input_tokens_seen": 1086912,
      "step": 1655
    },
    {
      "epoch": 0.870020964360587,
      "grad_norm": 0.5412030220031738,
      "learning_rate": 2.17374213836478e-05,
      "loss": 0.681,
      "num_input_tokens_seen": 1089696,
      "step": 1660
    },
    {
      "epoch": 0.8726415094339622,
      "grad_norm": 0.9703071713447571,
      "learning_rate": 2.1802935010482182e-05,
      "loss": 0.5289,
      "num_input_tokens_seen": 1092608,
      "step": 1665
    },
    {
      "epoch": 0.8752620545073375,
      "grad_norm": 0.3725551962852478,
      "learning_rate": 2.186844863731656e-05,
      "loss": 0.7744,
      "num_input_tokens_seen": 1095808,
      "step": 1670
    },
    {
      "epoch": 0.8778825995807128,
      "grad_norm": 0.4929797351360321,
      "learning_rate": 2.1933962264150945e-05,
      "loss": 0.6631,
      "num_input_tokens_seen": 1098816,
      "step": 1675
    },
    {
      "epoch": 0.8805031446540881,
      "grad_norm": 0.42914077639579773,
      "learning_rate": 2.1999475890985328e-05,
      "loss": 0.6448,
      "num_input_tokens_seen": 1103136,
      "step": 1680
    },
    {
      "epoch": 0.8831236897274634,
      "grad_norm": 0.45428207516670227,
      "learning_rate": 2.2064989517819707e-05,
      "loss": 0.5341,
      "num_input_tokens_seen": 1107872,
      "step": 1685
    },
    {
      "epoch": 0.8857442348008385,
      "grad_norm": 0.6143388748168945,
      "learning_rate": 2.2130503144654087e-05,
      "loss": 0.6872,
      "num_input_tokens_seen": 1111072,
      "step": 1690
    },
    {
      "epoch": 0.8883647798742138,
      "grad_norm": 1.4893620014190674,
      "learning_rate": 2.219601677148847e-05,
      "loss": 0.6674,
      "num_input_tokens_seen": 1113792,
      "step": 1695
    },
    {
      "epoch": 0.8909853249475891,
      "grad_norm": 0.8289670944213867,
      "learning_rate": 2.2261530398322853e-05,
      "loss": 0.6124,
      "num_input_tokens_seen": 1116384,
      "step": 1700
    },
    {
      "epoch": 0.8936058700209644,
      "grad_norm": 0.43715864419937134,
      "learning_rate": 2.2327044025157233e-05,
      "loss": 0.5576,
      "num_input_tokens_seen": 1119296,
      "step": 1705
    },
    {
      "epoch": 0.8962264150943396,
      "grad_norm": 0.47138169407844543,
      "learning_rate": 2.2392557651991616e-05,
      "loss": 0.6771,
      "num_input_tokens_seen": 1122720,
      "step": 1710
    },
    {
      "epoch": 0.8988469601677149,
      "grad_norm": 1.88077974319458,
      "learning_rate": 2.2458071278825996e-05,
      "loss": 0.7852,
      "num_input_tokens_seen": 1125088,
      "step": 1715
    },
    {
      "epoch": 0.9014675052410901,
      "grad_norm": 0.6047210693359375,
      "learning_rate": 2.252358490566038e-05,
      "loss": 0.7204,
      "num_input_tokens_seen": 1128288,
      "step": 1720
    },
    {
      "epoch": 0.9040880503144654,
      "grad_norm": 1.7054543495178223,
      "learning_rate": 2.258909853249476e-05,
      "loss": 0.7069,
      "num_input_tokens_seen": 1131040,
      "step": 1725
    },
    {
      "epoch": 0.9067085953878407,
      "grad_norm": 0.5555011630058289,
      "learning_rate": 2.265461215932914e-05,
      "loss": 0.5968,
      "num_input_tokens_seen": 1135136,
      "step": 1730
    },
    {
      "epoch": 0.9093291404612159,
      "grad_norm": 0.5053899884223938,
      "learning_rate": 2.2720125786163524e-05,
      "loss": 0.7413,
      "num_input_tokens_seen": 1139008,
      "step": 1735
    },
    {
      "epoch": 0.9119496855345912,
      "grad_norm": 0.5590983629226685,
      "learning_rate": 2.2785639412997904e-05,
      "loss": 0.5417,
      "num_input_tokens_seen": 1141920,
      "step": 1740
    },
    {
      "epoch": 0.9145702306079665,
      "grad_norm": 0.7543336749076843,
      "learning_rate": 2.2851153039832284e-05,
      "loss": 0.6172,
      "num_input_tokens_seen": 1144960,
      "step": 1745
    },
    {
      "epoch": 0.9171907756813418,
      "grad_norm": 0.9317483305931091,
      "learning_rate": 2.2916666666666667e-05,
      "loss": 0.6513,
      "num_input_tokens_seen": 1147712,
      "step": 1750
    },
    {
      "epoch": 0.9198113207547169,
      "grad_norm": 0.4906696379184723,
      "learning_rate": 2.298218029350105e-05,
      "loss": 0.6143,
      "num_input_tokens_seen": 1150528,
      "step": 1755
    },
    {
      "epoch": 0.9224318658280922,
      "grad_norm": 2.231757402420044,
      "learning_rate": 2.304769392033543e-05,
      "loss": 0.6818,
      "num_input_tokens_seen": 1154272,
      "step": 1760
    },
    {
      "epoch": 0.9250524109014675,
      "grad_norm": 1.0923089981079102,
      "learning_rate": 2.3113207547169813e-05,
      "loss": 0.596,
      "num_input_tokens_seen": 1157088,
      "step": 1765
    },
    {
      "epoch": 0.9276729559748428,
      "grad_norm": 0.4431854784488678,
      "learning_rate": 2.3178721174004196e-05,
      "loss": 0.7229,
      "num_input_tokens_seen": 1160192,
      "step": 1770
    },
    {
      "epoch": 0.9302935010482181,
      "grad_norm": 0.5254898071289062,
      "learning_rate": 2.3244234800838576e-05,
      "loss": 0.6043,
      "num_input_tokens_seen": 1163424,
      "step": 1775
    },
    {
      "epoch": 0.9329140461215933,
      "grad_norm": 0.4022844433784485,
      "learning_rate": 2.3309748427672955e-05,
      "loss": 0.653,
      "num_input_tokens_seen": 1167104,
      "step": 1780
    },
    {
      "epoch": 0.9355345911949685,
      "grad_norm": 1.1012533903121948,
      "learning_rate": 2.3375262054507338e-05,
      "loss": 0.604,
      "num_input_tokens_seen": 1170368,
      "step": 1785
    },
    {
      "epoch": 0.9381551362683438,
      "grad_norm": 0.5317506790161133,
      "learning_rate": 2.344077568134172e-05,
      "loss": 0.7018,
      "num_input_tokens_seen": 1173888,
      "step": 1790
    },
    {
      "epoch": 0.9407756813417191,
      "grad_norm": 0.4103237986564636,
      "learning_rate": 2.35062893081761e-05,
      "loss": 0.542,
      "num_input_tokens_seen": 1177408,
      "step": 1795
    },
    {
      "epoch": 0.9433962264150944,
      "grad_norm": 0.6778773665428162,
      "learning_rate": 2.3571802935010484e-05,
      "loss": 0.5519,
      "num_input_tokens_seen": 1181056,
      "step": 1800
    },
    {
      "epoch": 0.9460167714884696,
      "grad_norm": 0.603457510471344,
      "learning_rate": 2.3637316561844867e-05,
      "loss": 0.7689,
      "num_input_tokens_seen": 1184256,
      "step": 1805
    },
    {
      "epoch": 0.9486373165618449,
      "grad_norm": 0.8140215873718262,
      "learning_rate": 2.3702830188679247e-05,
      "loss": 0.5198,
      "num_input_tokens_seen": 1187168,
      "step": 1810
    },
    {
      "epoch": 0.9512578616352201,
      "grad_norm": 0.3597889840602875,
      "learning_rate": 2.3768343815513627e-05,
      "loss": 0.5154,
      "num_input_tokens_seen": 1190784,
      "step": 1815
    },
    {
      "epoch": 0.9538784067085954,
      "grad_norm": 0.6241923570632935,
      "learning_rate": 2.383385744234801e-05,
      "loss": 0.6839,
      "num_input_tokens_seen": 1193312,
      "step": 1820
    },
    {
      "epoch": 0.9564989517819706,
      "grad_norm": 0.48213183879852295,
      "learning_rate": 2.3899371069182393e-05,
      "loss": 0.6059,
      "num_input_tokens_seen": 1196608,
      "step": 1825
    },
    {
      "epoch": 0.9591194968553459,
      "grad_norm": 0.6304182410240173,
      "learning_rate": 2.3964884696016772e-05,
      "loss": 0.6567,
      "num_input_tokens_seen": 1200160,
      "step": 1830
    },
    {
      "epoch": 0.9617400419287212,
      "grad_norm": 0.9303752183914185,
      "learning_rate": 2.4030398322851155e-05,
      "loss": 0.5262,
      "num_input_tokens_seen": 1203040,
      "step": 1835
    },
    {
      "epoch": 0.9643605870020965,
      "grad_norm": 0.40780675411224365,
      "learning_rate": 2.4095911949685535e-05,
      "loss": 0.4398,
      "num_input_tokens_seen": 1205408,
      "step": 1840
    },
    {
      "epoch": 0.9669811320754716,
      "grad_norm": 1.8933926820755005,
      "learning_rate": 2.4161425576519918e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 1208896,
      "step": 1845
    },
    {
      "epoch": 0.9696016771488469,
      "grad_norm": 0.4080277681350708,
      "learning_rate": 2.4226939203354298e-05,
      "loss": 0.5623,
      "num_input_tokens_seen": 1212576,
      "step": 1850
    },
    {
      "epoch": 0.9722222222222222,
      "grad_norm": 0.39525383710861206,
      "learning_rate": 2.429245283018868e-05,
      "loss": 0.5551,
      "num_input_tokens_seen": 1215872,
      "step": 1855
    },
    {
      "epoch": 0.9748427672955975,
      "grad_norm": 0.6962201595306396,
      "learning_rate": 2.4357966457023064e-05,
      "loss": 0.5463,
      "num_input_tokens_seen": 1219968,
      "step": 1860
    },
    {
      "epoch": 0.9774633123689728,
      "grad_norm": 0.37134963274002075,
      "learning_rate": 2.4423480083857444e-05,
      "loss": 0.6082,
      "num_input_tokens_seen": 1222048,
      "step": 1865
    },
    {
      "epoch": 0.980083857442348,
      "grad_norm": 0.672056257724762,
      "learning_rate": 2.4488993710691823e-05,
      "loss": 0.6314,
      "num_input_tokens_seen": 1225120,
      "step": 1870
    },
    {
      "epoch": 0.9827044025157232,
      "grad_norm": 0.4278801679611206,
      "learning_rate": 2.4554507337526206e-05,
      "loss": 0.5823,
      "num_input_tokens_seen": 1229056,
      "step": 1875
    },
    {
      "epoch": 0.9853249475890985,
      "grad_norm": 1.0440847873687744,
      "learning_rate": 2.4620020964360586e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 1231776,
      "step": 1880
    },
    {
      "epoch": 0.9879454926624738,
      "grad_norm": 0.554691731929779,
      "learning_rate": 2.468553459119497e-05,
      "loss": 0.5674,
      "num_input_tokens_seen": 1235040,
      "step": 1885
    },
    {
      "epoch": 0.9905660377358491,
      "grad_norm": 0.6453990936279297,
      "learning_rate": 2.4751048218029352e-05,
      "loss": 0.5268,
      "num_input_tokens_seen": 1238048,
      "step": 1890
    },
    {
      "epoch": 0.9931865828092243,
      "grad_norm": 0.3971406817436218,
      "learning_rate": 2.4816561844863735e-05,
      "loss": 0.6059,
      "num_input_tokens_seen": 1241344,
      "step": 1895
    },
    {
      "epoch": 0.9958071278825996,
      "grad_norm": 0.6275870203971863,
      "learning_rate": 2.488207547169811e-05,
      "loss": 0.4554,
      "num_input_tokens_seen": 1243840,
      "step": 1900
    },
    {
      "epoch": 0.9984276729559748,
      "grad_norm": 0.43916407227516174,
      "learning_rate": 2.4947589098532495e-05,
      "loss": 0.6523,
      "num_input_tokens_seen": 1246816,
      "step": 1905
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.5840584635734558,
      "eval_runtime": 9.7323,
      "eval_samples_per_second": 87.133,
      "eval_steps_per_second": 21.783,
      "num_input_tokens_seen": 1248304,
      "step": 1908
    },
    {
      "epoch": 1.00104821802935,
      "grad_norm": 0.6606485843658447,
      "learning_rate": 2.5013102725366878e-05,
      "loss": 0.5836,
      "num_input_tokens_seen": 1249200,
      "step": 1910
    },
    {
      "epoch": 1.0036687631027255,
      "grad_norm": 1.2158397436141968,
      "learning_rate": 2.507861635220126e-05,
      "loss": 0.6168,
      "num_input_tokens_seen": 1251664,
      "step": 1915
    },
    {
      "epoch": 1.0062893081761006,
      "grad_norm": 0.34211447834968567,
      "learning_rate": 2.514412997903564e-05,
      "loss": 0.502,
      "num_input_tokens_seen": 1254928,
      "step": 1920
    },
    {
      "epoch": 1.0089098532494758,
      "grad_norm": 1.0797604322433472,
      "learning_rate": 2.520964360587002e-05,
      "loss": 0.474,
      "num_input_tokens_seen": 1257744,
      "step": 1925
    },
    {
      "epoch": 1.0115303983228512,
      "grad_norm": 0.8401476144790649,
      "learning_rate": 2.5275157232704407e-05,
      "loss": 0.5192,
      "num_input_tokens_seen": 1261904,
      "step": 1930
    },
    {
      "epoch": 1.0141509433962264,
      "grad_norm": 0.4762633442878723,
      "learning_rate": 2.5340670859538783e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 1264432,
      "step": 1935
    },
    {
      "epoch": 1.0167714884696017,
      "grad_norm": 0.9594195485115051,
      "learning_rate": 2.540618448637317e-05,
      "loss": 0.6495,
      "num_input_tokens_seen": 1268400,
      "step": 1940
    },
    {
      "epoch": 1.019392033542977,
      "grad_norm": 0.5448996424674988,
      "learning_rate": 2.547169811320755e-05,
      "loss": 0.5089,
      "num_input_tokens_seen": 1272752,
      "step": 1945
    },
    {
      "epoch": 1.0220125786163523,
      "grad_norm": 0.5377333164215088,
      "learning_rate": 2.553721174004193e-05,
      "loss": 0.6505,
      "num_input_tokens_seen": 1275568,
      "step": 1950
    },
    {
      "epoch": 1.0246331236897275,
      "grad_norm": 0.4580249786376953,
      "learning_rate": 2.5602725366876312e-05,
      "loss": 0.6233,
      "num_input_tokens_seen": 1278160,
      "step": 1955
    },
    {
      "epoch": 1.0272536687631026,
      "grad_norm": 0.6271491646766663,
      "learning_rate": 2.566823899371069e-05,
      "loss": 0.5561,
      "num_input_tokens_seen": 1281424,
      "step": 1960
    },
    {
      "epoch": 1.029874213836478,
      "grad_norm": 0.6026850938796997,
      "learning_rate": 2.5733752620545075e-05,
      "loss": 0.5559,
      "num_input_tokens_seen": 1283920,
      "step": 1965
    },
    {
      "epoch": 1.0324947589098532,
      "grad_norm": 0.754227876663208,
      "learning_rate": 2.5799266247379454e-05,
      "loss": 0.5203,
      "num_input_tokens_seen": 1287920,
      "step": 1970
    },
    {
      "epoch": 1.0351153039832286,
      "grad_norm": 1.7846163511276245,
      "learning_rate": 2.586477987421384e-05,
      "loss": 0.5973,
      "num_input_tokens_seen": 1290960,
      "step": 1975
    },
    {
      "epoch": 1.0377358490566038,
      "grad_norm": 0.4974835515022278,
      "learning_rate": 2.593029350104822e-05,
      "loss": 0.4491,
      "num_input_tokens_seen": 1294544,
      "step": 1980
    },
    {
      "epoch": 1.040356394129979,
      "grad_norm": 0.3582246005535126,
      "learning_rate": 2.59958071278826e-05,
      "loss": 0.7018,
      "num_input_tokens_seen": 1298160,
      "step": 1985
    },
    {
      "epoch": 1.0429769392033543,
      "grad_norm": 0.548183023929596,
      "learning_rate": 2.6061320754716983e-05,
      "loss": 0.5816,
      "num_input_tokens_seen": 1300368,
      "step": 1990
    },
    {
      "epoch": 1.0455974842767295,
      "grad_norm": 0.27929338812828064,
      "learning_rate": 2.6126834381551363e-05,
      "loss": 0.4723,
      "num_input_tokens_seen": 1307184,
      "step": 1995
    },
    {
      "epoch": 1.0482180293501049,
      "grad_norm": 0.6546352505683899,
      "learning_rate": 2.6192348008385746e-05,
      "loss": 0.8861,
      "num_input_tokens_seen": 1309840,
      "step": 2000
    },
    {
      "epoch": 1.05083857442348,
      "grad_norm": 0.7085390686988831,
      "learning_rate": 2.6257861635220126e-05,
      "loss": 0.52,
      "num_input_tokens_seen": 1313648,
      "step": 2005
    },
    {
      "epoch": 1.0534591194968554,
      "grad_norm": 0.5142965912818909,
      "learning_rate": 2.6323375262054505e-05,
      "loss": 0.5325,
      "num_input_tokens_seen": 1316496,
      "step": 2010
    },
    {
      "epoch": 1.0560796645702306,
      "grad_norm": 1.2046747207641602,
      "learning_rate": 2.6388888888888892e-05,
      "loss": 0.5755,
      "num_input_tokens_seen": 1319952,
      "step": 2015
    },
    {
      "epoch": 1.0587002096436058,
      "grad_norm": 0.44213345646858215,
      "learning_rate": 2.645440251572327e-05,
      "loss": 0.5177,
      "num_input_tokens_seen": 1323344,
      "step": 2020
    },
    {
      "epoch": 1.0613207547169812,
      "grad_norm": 0.5797549486160278,
      "learning_rate": 2.6519916142557655e-05,
      "loss": 0.618,
      "num_input_tokens_seen": 1326704,
      "step": 2025
    },
    {
      "epoch": 1.0639412997903563,
      "grad_norm": 0.4801445007324219,
      "learning_rate": 2.6585429769392034e-05,
      "loss": 0.5467,
      "num_input_tokens_seen": 1330128,
      "step": 2030
    },
    {
      "epoch": 1.0665618448637317,
      "grad_norm": 0.5947191715240479,
      "learning_rate": 2.6650943396226417e-05,
      "loss": 0.492,
      "num_input_tokens_seen": 1333040,
      "step": 2035
    },
    {
      "epoch": 1.069182389937107,
      "grad_norm": 0.5451018810272217,
      "learning_rate": 2.6716457023060797e-05,
      "loss": 0.5648,
      "num_input_tokens_seen": 1335504,
      "step": 2040
    },
    {
      "epoch": 1.0718029350104823,
      "grad_norm": 0.8190847039222717,
      "learning_rate": 2.6781970649895177e-05,
      "loss": 0.5834,
      "num_input_tokens_seen": 1339280,
      "step": 2045
    },
    {
      "epoch": 1.0744234800838575,
      "grad_norm": 0.7532840371131897,
      "learning_rate": 2.6847484276729563e-05,
      "loss": 0.6217,
      "num_input_tokens_seen": 1342576,
      "step": 2050
    },
    {
      "epoch": 1.0770440251572326,
      "grad_norm": 0.5581588745117188,
      "learning_rate": 2.6912997903563943e-05,
      "loss": 0.5272,
      "num_input_tokens_seen": 1346256,
      "step": 2055
    },
    {
      "epoch": 1.079664570230608,
      "grad_norm": 0.6396252512931824,
      "learning_rate": 2.6978511530398326e-05,
      "loss": 0.6655,
      "num_input_tokens_seen": 1349008,
      "step": 2060
    },
    {
      "epoch": 1.0822851153039832,
      "grad_norm": 1.0858092308044434,
      "learning_rate": 2.7044025157232706e-05,
      "loss": 0.5936,
      "num_input_tokens_seen": 1353552,
      "step": 2065
    },
    {
      "epoch": 1.0849056603773586,
      "grad_norm": 1.0294277667999268,
      "learning_rate": 2.710953878406709e-05,
      "loss": 0.5392,
      "num_input_tokens_seen": 1356400,
      "step": 2070
    },
    {
      "epoch": 1.0875262054507338,
      "grad_norm": 0.7925314903259277,
      "learning_rate": 2.717505241090147e-05,
      "loss": 0.5908,
      "num_input_tokens_seen": 1361008,
      "step": 2075
    },
    {
      "epoch": 1.090146750524109,
      "grad_norm": 1.4446648359298706,
      "learning_rate": 2.7240566037735848e-05,
      "loss": 0.6693,
      "num_input_tokens_seen": 1363184,
      "step": 2080
    },
    {
      "epoch": 1.0927672955974843,
      "grad_norm": 1.1559408903121948,
      "learning_rate": 2.7306079664570235e-05,
      "loss": 0.5611,
      "num_input_tokens_seen": 1365872,
      "step": 2085
    },
    {
      "epoch": 1.0953878406708595,
      "grad_norm": 0.5468798875808716,
      "learning_rate": 2.737159329140461e-05,
      "loss": 0.6276,
      "num_input_tokens_seen": 1369168,
      "step": 2090
    },
    {
      "epoch": 1.0980083857442349,
      "grad_norm": 0.4694921374320984,
      "learning_rate": 2.7437106918238997e-05,
      "loss": 0.5349,
      "num_input_tokens_seen": 1372208,
      "step": 2095
    },
    {
      "epoch": 1.10062893081761,
      "grad_norm": 0.5734136700630188,
      "learning_rate": 2.7502620545073377e-05,
      "loss": 0.4399,
      "num_input_tokens_seen": 1377360,
      "step": 2100
    },
    {
      "epoch": 1.1032494758909852,
      "grad_norm": 0.6910477876663208,
      "learning_rate": 2.7568134171907757e-05,
      "loss": 0.4992,
      "num_input_tokens_seen": 1380400,
      "step": 2105
    },
    {
      "epoch": 1.1058700209643606,
      "grad_norm": 0.8054786324501038,
      "learning_rate": 2.763364779874214e-05,
      "loss": 0.5533,
      "num_input_tokens_seen": 1383056,
      "step": 2110
    },
    {
      "epoch": 1.1084905660377358,
      "grad_norm": 0.5916509628295898,
      "learning_rate": 2.769916142557652e-05,
      "loss": 0.6486,
      "num_input_tokens_seen": 1386160,
      "step": 2115
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 0.6389366388320923,
      "learning_rate": 2.7764675052410906e-05,
      "loss": 0.5046,
      "num_input_tokens_seen": 1389232,
      "step": 2120
    },
    {
      "epoch": 1.1137316561844863,
      "grad_norm": 0.4738309383392334,
      "learning_rate": 2.7830188679245282e-05,
      "loss": 0.665,
      "num_input_tokens_seen": 1393328,
      "step": 2125
    },
    {
      "epoch": 1.1163522012578617,
      "grad_norm": 0.49418917298316956,
      "learning_rate": 2.789570230607967e-05,
      "loss": 0.4791,
      "num_input_tokens_seen": 1396816,
      "step": 2130
    },
    {
      "epoch": 1.118972746331237,
      "grad_norm": 0.3958026170730591,
      "learning_rate": 2.796121593291405e-05,
      "loss": 0.5025,
      "num_input_tokens_seen": 1400176,
      "step": 2135
    },
    {
      "epoch": 1.121593291404612,
      "grad_norm": 0.5744937062263489,
      "learning_rate": 2.8026729559748428e-05,
      "loss": 0.6662,
      "num_input_tokens_seen": 1403408,
      "step": 2140
    },
    {
      "epoch": 1.1242138364779874,
      "grad_norm": 0.4886334538459778,
      "learning_rate": 2.809224318658281e-05,
      "loss": 0.5469,
      "num_input_tokens_seen": 1406448,
      "step": 2145
    },
    {
      "epoch": 1.1268343815513626,
      "grad_norm": 1.5042985677719116,
      "learning_rate": 2.815775681341719e-05,
      "loss": 0.6144,
      "num_input_tokens_seen": 1409584,
      "step": 2150
    },
    {
      "epoch": 1.129454926624738,
      "grad_norm": 0.6764423251152039,
      "learning_rate": 2.8223270440251577e-05,
      "loss": 0.5215,
      "num_input_tokens_seen": 1413872,
      "step": 2155
    },
    {
      "epoch": 1.1320754716981132,
      "grad_norm": 0.5423761606216431,
      "learning_rate": 2.8288784067085954e-05,
      "loss": 0.7403,
      "num_input_tokens_seen": 1416528,
      "step": 2160
    },
    {
      "epoch": 1.1346960167714886,
      "grad_norm": 0.7160435318946838,
      "learning_rate": 2.8354297693920333e-05,
      "loss": 0.5465,
      "num_input_tokens_seen": 1418896,
      "step": 2165
    },
    {
      "epoch": 1.1373165618448637,
      "grad_norm": 0.7401747703552246,
      "learning_rate": 2.841981132075472e-05,
      "loss": 0.6061,
      "num_input_tokens_seen": 1421712,
      "step": 2170
    },
    {
      "epoch": 1.139937106918239,
      "grad_norm": 0.49508941173553467,
      "learning_rate": 2.84853249475891e-05,
      "loss": 0.5335,
      "num_input_tokens_seen": 1425072,
      "step": 2175
    },
    {
      "epoch": 1.1425576519916143,
      "grad_norm": 0.3371657133102417,
      "learning_rate": 2.8550838574423482e-05,
      "loss": 0.7107,
      "num_input_tokens_seen": 1428848,
      "step": 2180
    },
    {
      "epoch": 1.1451781970649895,
      "grad_norm": 0.7862048745155334,
      "learning_rate": 2.8616352201257862e-05,
      "loss": 0.6093,
      "num_input_tokens_seen": 1432432,
      "step": 2185
    },
    {
      "epoch": 1.1477987421383649,
      "grad_norm": 0.47624900937080383,
      "learning_rate": 2.868186582809225e-05,
      "loss": 0.7142,
      "num_input_tokens_seen": 1436048,
      "step": 2190
    },
    {
      "epoch": 1.15041928721174,
      "grad_norm": 0.8774187564849854,
      "learning_rate": 2.8747379454926625e-05,
      "loss": 0.7227,
      "num_input_tokens_seen": 1440144,
      "step": 2195
    },
    {
      "epoch": 1.1530398322851152,
      "grad_norm": 0.579868495464325,
      "learning_rate": 2.8812893081761005e-05,
      "loss": 0.6514,
      "num_input_tokens_seen": 1443760,
      "step": 2200
    },
    {
      "epoch": 1.1556603773584906,
      "grad_norm": 0.5374531149864197,
      "learning_rate": 2.887840670859539e-05,
      "loss": 0.8422,
      "num_input_tokens_seen": 1446864,
      "step": 2205
    },
    {
      "epoch": 1.1582809224318658,
      "grad_norm": 0.7106645703315735,
      "learning_rate": 2.894392033542977e-05,
      "loss": 0.3808,
      "num_input_tokens_seen": 1449456,
      "step": 2210
    },
    {
      "epoch": 1.1609014675052411,
      "grad_norm": 0.594680666923523,
      "learning_rate": 2.9009433962264154e-05,
      "loss": 0.6432,
      "num_input_tokens_seen": 1452944,
      "step": 2215
    },
    {
      "epoch": 1.1635220125786163,
      "grad_norm": 0.324680358171463,
      "learning_rate": 2.9074947589098533e-05,
      "loss": 0.5333,
      "num_input_tokens_seen": 1455952,
      "step": 2220
    },
    {
      "epoch": 1.1661425576519917,
      "grad_norm": 0.5021088719367981,
      "learning_rate": 2.9140461215932913e-05,
      "loss": 0.6884,
      "num_input_tokens_seen": 1459376,
      "step": 2225
    },
    {
      "epoch": 1.1687631027253669,
      "grad_norm": 0.560049295425415,
      "learning_rate": 2.9205974842767296e-05,
      "loss": 0.5356,
      "num_input_tokens_seen": 1462352,
      "step": 2230
    },
    {
      "epoch": 1.171383647798742,
      "grad_norm": 0.6400923132896423,
      "learning_rate": 2.9271488469601676e-05,
      "loss": 0.6474,
      "num_input_tokens_seen": 1465104,
      "step": 2235
    },
    {
      "epoch": 1.1740041928721174,
      "grad_norm": 0.4037795960903168,
      "learning_rate": 2.9337002096436062e-05,
      "loss": 0.6259,
      "num_input_tokens_seen": 1468624,
      "step": 2240
    },
    {
      "epoch": 1.1766247379454926,
      "grad_norm": 0.5376235246658325,
      "learning_rate": 2.9402515723270442e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 1475184,
      "step": 2245
    },
    {
      "epoch": 1.179245283018868,
      "grad_norm": 0.27118128538131714,
      "learning_rate": 2.9468029350104825e-05,
      "loss": 0.4619,
      "num_input_tokens_seen": 1478640,
      "step": 2250
    },
    {
      "epoch": 1.1818658280922432,
      "grad_norm": 0.4761984646320343,
      "learning_rate": 2.9533542976939205e-05,
      "loss": 0.63,
      "num_input_tokens_seen": 1481552,
      "step": 2255
    },
    {
      "epoch": 1.1844863731656186,
      "grad_norm": 0.8309112191200256,
      "learning_rate": 2.9599056603773584e-05,
      "loss": 0.5587,
      "num_input_tokens_seen": 1485104,
      "step": 2260
    },
    {
      "epoch": 1.1871069182389937,
      "grad_norm": 0.5139539837837219,
      "learning_rate": 2.9664570230607968e-05,
      "loss": 0.6486,
      "num_input_tokens_seen": 1487728,
      "step": 2265
    },
    {
      "epoch": 1.189727463312369,
      "grad_norm": 0.4605681002140045,
      "learning_rate": 2.9730083857442347e-05,
      "loss": 0.4954,
      "num_input_tokens_seen": 1491312,
      "step": 2270
    },
    {
      "epoch": 1.1923480083857443,
      "grad_norm": 0.6950439810752869,
      "learning_rate": 2.9795597484276734e-05,
      "loss": 0.6381,
      "num_input_tokens_seen": 1494192,
      "step": 2275
    },
    {
      "epoch": 1.1949685534591195,
      "grad_norm": 0.6353591680526733,
      "learning_rate": 2.9861111111111113e-05,
      "loss": 0.7162,
      "num_input_tokens_seen": 1497168,
      "step": 2280
    },
    {
      "epoch": 1.1975890985324948,
      "grad_norm": 0.49460119009017944,
      "learning_rate": 2.9926624737945496e-05,
      "loss": 0.5573,
      "num_input_tokens_seen": 1501936,
      "step": 2285
    },
    {
      "epoch": 1.20020964360587,
      "grad_norm": 0.41559964418411255,
      "learning_rate": 2.9992138364779876e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 1505040,
      "step": 2290
    },
    {
      "epoch": 1.2028301886792452,
      "grad_norm": 0.9528219699859619,
      "learning_rate": 3.0057651991614256e-05,
      "loss": 0.7057,
      "num_input_tokens_seen": 1507856,
      "step": 2295
    },
    {
      "epoch": 1.2054507337526206,
      "grad_norm": 0.6479001641273499,
      "learning_rate": 3.012316561844864e-05,
      "loss": 0.5637,
      "num_input_tokens_seen": 1510768,
      "step": 2300
    },
    {
      "epoch": 1.2080712788259957,
      "grad_norm": 0.6436383724212646,
      "learning_rate": 3.018867924528302e-05,
      "loss": 0.6821,
      "num_input_tokens_seen": 1513808,
      "step": 2305
    },
    {
      "epoch": 1.2106918238993711,
      "grad_norm": 0.4581257998943329,
      "learning_rate": 3.0254192872117405e-05,
      "loss": 0.5391,
      "num_input_tokens_seen": 1517392,
      "step": 2310
    },
    {
      "epoch": 1.2133123689727463,
      "grad_norm": 0.35487857460975647,
      "learning_rate": 3.0319706498951785e-05,
      "loss": 0.5785,
      "num_input_tokens_seen": 1521648,
      "step": 2315
    },
    {
      "epoch": 1.2159329140461215,
      "grad_norm": 0.5633171200752258,
      "learning_rate": 3.038522012578616e-05,
      "loss": 0.4906,
      "num_input_tokens_seen": 1524496,
      "step": 2320
    },
    {
      "epoch": 1.2185534591194969,
      "grad_norm": 0.5190237164497375,
      "learning_rate": 3.0450733752620547e-05,
      "loss": 0.539,
      "num_input_tokens_seen": 1528144,
      "step": 2325
    },
    {
      "epoch": 1.221174004192872,
      "grad_norm": 0.6242536902427673,
      "learning_rate": 3.0516247379454927e-05,
      "loss": 0.5356,
      "num_input_tokens_seen": 1531664,
      "step": 2330
    },
    {
      "epoch": 1.2237945492662474,
      "grad_norm": 0.6050558686256409,
      "learning_rate": 3.0581761006289314e-05,
      "loss": 0.649,
      "num_input_tokens_seen": 1535408,
      "step": 2335
    },
    {
      "epoch": 1.2264150943396226,
      "grad_norm": 0.42427489161491394,
      "learning_rate": 3.064727463312369e-05,
      "loss": 0.574,
      "num_input_tokens_seen": 1538448,
      "step": 2340
    },
    {
      "epoch": 1.229035639412998,
      "grad_norm": 0.526871919631958,
      "learning_rate": 3.071278825995807e-05,
      "loss": 0.6188,
      "num_input_tokens_seen": 1541968,
      "step": 2345
    },
    {
      "epoch": 1.2316561844863732,
      "grad_norm": 0.38725200295448303,
      "learning_rate": 3.0778301886792456e-05,
      "loss": 0.5973,
      "num_input_tokens_seen": 1544560,
      "step": 2350
    },
    {
      "epoch": 1.2342767295597485,
      "grad_norm": 0.6963697671890259,
      "learning_rate": 3.084381551362683e-05,
      "loss": 0.5775,
      "num_input_tokens_seen": 1548080,
      "step": 2355
    },
    {
      "epoch": 1.2368972746331237,
      "grad_norm": 0.7571613788604736,
      "learning_rate": 3.0909329140461215e-05,
      "loss": 0.4871,
      "num_input_tokens_seen": 1551472,
      "step": 2360
    },
    {
      "epoch": 1.2395178197064989,
      "grad_norm": 0.4055423438549042,
      "learning_rate": 3.09748427672956e-05,
      "loss": 0.6556,
      "num_input_tokens_seen": 1553904,
      "step": 2365
    },
    {
      "epoch": 1.2421383647798743,
      "grad_norm": 0.45292335748672485,
      "learning_rate": 3.104035639412998e-05,
      "loss": 0.5079,
      "num_input_tokens_seen": 1558352,
      "step": 2370
    },
    {
      "epoch": 1.2447589098532494,
      "grad_norm": 0.6336278319358826,
      "learning_rate": 3.1105870020964365e-05,
      "loss": 0.4328,
      "num_input_tokens_seen": 1561616,
      "step": 2375
    },
    {
      "epoch": 1.2473794549266248,
      "grad_norm": 0.4950680732727051,
      "learning_rate": 3.117138364779874e-05,
      "loss": 0.434,
      "num_input_tokens_seen": 1566000,
      "step": 2380
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.5517417788505554,
      "learning_rate": 3.1236897274633124e-05,
      "loss": 0.5704,
      "num_input_tokens_seen": 1569776,
      "step": 2385
    },
    {
      "epoch": 1.2526205450733752,
      "grad_norm": 0.8537105917930603,
      "learning_rate": 3.130241090146751e-05,
      "loss": 0.6594,
      "num_input_tokens_seen": 1572112,
      "step": 2390
    },
    {
      "epoch": 1.2552410901467506,
      "grad_norm": 0.47332292795181274,
      "learning_rate": 3.136792452830189e-05,
      "loss": 0.5747,
      "num_input_tokens_seen": 1575312,
      "step": 2395
    },
    {
      "epoch": 1.2578616352201257,
      "grad_norm": 0.4547702968120575,
      "learning_rate": 3.1433438155136266e-05,
      "loss": 0.6204,
      "num_input_tokens_seen": 1578256,
      "step": 2400
    },
    {
      "epoch": 1.2604821802935011,
      "grad_norm": 0.5289472937583923,
      "learning_rate": 3.1498951781970656e-05,
      "loss": 0.7444,
      "num_input_tokens_seen": 1581328,
      "step": 2405
    },
    {
      "epoch": 1.2631027253668763,
      "grad_norm": 0.5441675186157227,
      "learning_rate": 3.156446540880503e-05,
      "loss": 0.6218,
      "num_input_tokens_seen": 1583952,
      "step": 2410
    },
    {
      "epoch": 1.2657232704402515,
      "grad_norm": 0.8311269283294678,
      "learning_rate": 3.162997903563941e-05,
      "loss": 0.5824,
      "num_input_tokens_seen": 1587472,
      "step": 2415
    },
    {
      "epoch": 1.2683438155136268,
      "grad_norm": 0.5005359053611755,
      "learning_rate": 3.16954926624738e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 1590384,
      "step": 2420
    },
    {
      "epoch": 1.270964360587002,
      "grad_norm": 0.5782462358474731,
      "learning_rate": 3.1761006289308175e-05,
      "loss": 0.5149,
      "num_input_tokens_seen": 1593264,
      "step": 2425
    },
    {
      "epoch": 1.2735849056603774,
      "grad_norm": 0.6015874147415161,
      "learning_rate": 3.182651991614256e-05,
      "loss": 0.5039,
      "num_input_tokens_seen": 1597680,
      "step": 2430
    },
    {
      "epoch": 1.2762054507337526,
      "grad_norm": 0.8269503712654114,
      "learning_rate": 3.189203354297694e-05,
      "loss": 0.5794,
      "num_input_tokens_seen": 1600912,
      "step": 2435
    },
    {
      "epoch": 1.2788259958071277,
      "grad_norm": 0.6123737096786499,
      "learning_rate": 3.1957547169811324e-05,
      "loss": 0.5082,
      "num_input_tokens_seen": 1603472,
      "step": 2440
    },
    {
      "epoch": 1.2814465408805031,
      "grad_norm": 1.407930850982666,
      "learning_rate": 3.202306079664571e-05,
      "loss": 0.5435,
      "num_input_tokens_seen": 1606384,
      "step": 2445
    },
    {
      "epoch": 1.2840670859538785,
      "grad_norm": 0.47395163774490356,
      "learning_rate": 3.2088574423480084e-05,
      "loss": 0.5471,
      "num_input_tokens_seen": 1609360,
      "step": 2450
    },
    {
      "epoch": 1.2866876310272537,
      "grad_norm": 1.122003436088562,
      "learning_rate": 3.215408805031447e-05,
      "loss": 0.5755,
      "num_input_tokens_seen": 1612336,
      "step": 2455
    },
    {
      "epoch": 1.2893081761006289,
      "grad_norm": 0.42012280225753784,
      "learning_rate": 3.221960167714885e-05,
      "loss": 0.4715,
      "num_input_tokens_seen": 1616048,
      "step": 2460
    },
    {
      "epoch": 1.2919287211740043,
      "grad_norm": 0.9056195020675659,
      "learning_rate": 3.228511530398323e-05,
      "loss": 0.4678,
      "num_input_tokens_seen": 1618960,
      "step": 2465
    },
    {
      "epoch": 1.2945492662473794,
      "grad_norm": 0.34666967391967773,
      "learning_rate": 3.235062893081761e-05,
      "loss": 0.4159,
      "num_input_tokens_seen": 1622544,
      "step": 2470
    },
    {
      "epoch": 1.2971698113207548,
      "grad_norm": 0.4146490693092346,
      "learning_rate": 3.241614255765199e-05,
      "loss": 0.5168,
      "num_input_tokens_seen": 1625872,
      "step": 2475
    },
    {
      "epoch": 1.29979035639413,
      "grad_norm": 0.38569217920303345,
      "learning_rate": 3.2481656184486375e-05,
      "loss": 0.5444,
      "num_input_tokens_seen": 1629456,
      "step": 2480
    },
    {
      "epoch": 1.3024109014675052,
      "grad_norm": 0.3460215628147125,
      "learning_rate": 3.254716981132075e-05,
      "loss": 0.6934,
      "num_input_tokens_seen": 1632432,
      "step": 2485
    },
    {
      "epoch": 1.3050314465408805,
      "grad_norm": 0.4276449680328369,
      "learning_rate": 3.261268343815514e-05,
      "loss": 0.5214,
      "num_input_tokens_seen": 1636080,
      "step": 2490
    },
    {
      "epoch": 1.3076519916142557,
      "grad_norm": 0.388629287481308,
      "learning_rate": 3.267819706498952e-05,
      "loss": 0.5609,
      "num_input_tokens_seen": 1638960,
      "step": 2495
    },
    {
      "epoch": 1.310272536687631,
      "grad_norm": 0.23330791294574738,
      "learning_rate": 3.27437106918239e-05,
      "loss": 0.535,
      "num_input_tokens_seen": 1643120,
      "step": 2500
    },
    {
      "epoch": 1.3128930817610063,
      "grad_norm": 0.6152562499046326,
      "learning_rate": 3.2809224318658284e-05,
      "loss": 0.4541,
      "num_input_tokens_seen": 1646768,
      "step": 2505
    },
    {
      "epoch": 1.3155136268343814,
      "grad_norm": 0.4153400659561157,
      "learning_rate": 3.287473794549266e-05,
      "loss": 0.5311,
      "num_input_tokens_seen": 1650256,
      "step": 2510
    },
    {
      "epoch": 1.3181341719077568,
      "grad_norm": 0.5374495387077332,
      "learning_rate": 3.294025157232704e-05,
      "loss": 0.6012,
      "num_input_tokens_seen": 1652912,
      "step": 2515
    },
    {
      "epoch": 1.320754716981132,
      "grad_norm": 0.3062704801559448,
      "learning_rate": 3.3005765199161426e-05,
      "loss": 0.5437,
      "num_input_tokens_seen": 1656432,
      "step": 2520
    },
    {
      "epoch": 1.3233752620545074,
      "grad_norm": 0.5008264780044556,
      "learning_rate": 3.307127882599581e-05,
      "loss": 0.4594,
      "num_input_tokens_seen": 1659472,
      "step": 2525
    },
    {
      "epoch": 1.3259958071278826,
      "grad_norm": 1.067449927330017,
      "learning_rate": 3.313679245283019e-05,
      "loss": 0.5648,
      "num_input_tokens_seen": 1662480,
      "step": 2530
    },
    {
      "epoch": 1.3286163522012577,
      "grad_norm": 0.5790795087814331,
      "learning_rate": 3.320230607966457e-05,
      "loss": 0.5861,
      "num_input_tokens_seen": 1665424,
      "step": 2535
    },
    {
      "epoch": 1.3312368972746331,
      "grad_norm": 0.45017656683921814,
      "learning_rate": 3.326781970649895e-05,
      "loss": 0.5994,
      "num_input_tokens_seen": 1667760,
      "step": 2540
    },
    {
      "epoch": 1.3338574423480085,
      "grad_norm": 0.535036563873291,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.4736,
      "num_input_tokens_seen": 1670800,
      "step": 2545
    },
    {
      "epoch": 1.3364779874213837,
      "grad_norm": 0.4800187647342682,
      "learning_rate": 3.339884696016772e-05,
      "loss": 0.6017,
      "num_input_tokens_seen": 1675760,
      "step": 2550
    },
    {
      "epoch": 1.3390985324947589,
      "grad_norm": 1.4444485902786255,
      "learning_rate": 3.3464360587002094e-05,
      "loss": 0.6079,
      "num_input_tokens_seen": 1677936,
      "step": 2555
    },
    {
      "epoch": 1.3417190775681342,
      "grad_norm": 0.377248078584671,
      "learning_rate": 3.3529874213836484e-05,
      "loss": 0.6699,
      "num_input_tokens_seen": 1685008,
      "step": 2560
    },
    {
      "epoch": 1.3443396226415094,
      "grad_norm": 0.5573608875274658,
      "learning_rate": 3.359538784067086e-05,
      "loss": 0.7223,
      "num_input_tokens_seen": 1688240,
      "step": 2565
    },
    {
      "epoch": 1.3469601677148848,
      "grad_norm": 0.5471317768096924,
      "learning_rate": 3.3660901467505244e-05,
      "loss": 0.7342,
      "num_input_tokens_seen": 1690832,
      "step": 2570
    },
    {
      "epoch": 1.34958071278826,
      "grad_norm": 0.534151554107666,
      "learning_rate": 3.3726415094339627e-05,
      "loss": 0.6744,
      "num_input_tokens_seen": 1695248,
      "step": 2575
    },
    {
      "epoch": 1.3522012578616351,
      "grad_norm": 0.7248173356056213,
      "learning_rate": 3.3791928721174e-05,
      "loss": 0.576,
      "num_input_tokens_seen": 1698896,
      "step": 2580
    },
    {
      "epoch": 1.3548218029350105,
      "grad_norm": 0.7537935972213745,
      "learning_rate": 3.3857442348008386e-05,
      "loss": 0.6965,
      "num_input_tokens_seen": 1701296,
      "step": 2585
    },
    {
      "epoch": 1.3574423480083857,
      "grad_norm": 0.7935356497764587,
      "learning_rate": 3.392295597484277e-05,
      "loss": 0.4071,
      "num_input_tokens_seen": 1705136,
      "step": 2590
    },
    {
      "epoch": 1.360062893081761,
      "grad_norm": 0.46900641918182373,
      "learning_rate": 3.398846960167715e-05,
      "loss": 0.5238,
      "num_input_tokens_seen": 1708016,
      "step": 2595
    },
    {
      "epoch": 1.3626834381551363,
      "grad_norm": 2.809082269668579,
      "learning_rate": 3.4053983228511535e-05,
      "loss": 0.7373,
      "num_input_tokens_seen": 1710864,
      "step": 2600
    },
    {
      "epoch": 1.3653039832285114,
      "grad_norm": 0.5377600789070129,
      "learning_rate": 3.411949685534591e-05,
      "loss": 0.487,
      "num_input_tokens_seen": 1713552,
      "step": 2605
    },
    {
      "epoch": 1.3679245283018868,
      "grad_norm": 0.5637503862380981,
      "learning_rate": 3.4185010482180295e-05,
      "loss": 0.4753,
      "num_input_tokens_seen": 1716528,
      "step": 2610
    },
    {
      "epoch": 1.370545073375262,
      "grad_norm": 1.5007275342941284,
      "learning_rate": 3.425052410901468e-05,
      "loss": 0.5333,
      "num_input_tokens_seen": 1719216,
      "step": 2615
    },
    {
      "epoch": 1.3731656184486374,
      "grad_norm": 0.4583808481693268,
      "learning_rate": 3.431603773584906e-05,
      "loss": 0.5752,
      "num_input_tokens_seen": 1722384,
      "step": 2620
    },
    {
      "epoch": 1.3757861635220126,
      "grad_norm": 0.42789044976234436,
      "learning_rate": 3.438155136268344e-05,
      "loss": 0.4768,
      "num_input_tokens_seen": 1725104,
      "step": 2625
    },
    {
      "epoch": 1.3784067085953877,
      "grad_norm": 0.7439911961555481,
      "learning_rate": 3.444706498951782e-05,
      "loss": 0.4754,
      "num_input_tokens_seen": 1728112,
      "step": 2630
    },
    {
      "epoch": 1.381027253668763,
      "grad_norm": 0.5652426481246948,
      "learning_rate": 3.45125786163522e-05,
      "loss": 0.6115,
      "num_input_tokens_seen": 1731536,
      "step": 2635
    },
    {
      "epoch": 1.3836477987421385,
      "grad_norm": 0.4301947355270386,
      "learning_rate": 3.457809224318658e-05,
      "loss": 0.6161,
      "num_input_tokens_seen": 1734896,
      "step": 2640
    },
    {
      "epoch": 1.3862683438155137,
      "grad_norm": 0.6778525710105896,
      "learning_rate": 3.464360587002097e-05,
      "loss": 0.5852,
      "num_input_tokens_seen": 1737584,
      "step": 2645
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 0.4971335828304291,
      "learning_rate": 3.4709119496855346e-05,
      "loss": 0.6388,
      "num_input_tokens_seen": 1740720,
      "step": 2650
    },
    {
      "epoch": 1.3915094339622642,
      "grad_norm": 0.5425771474838257,
      "learning_rate": 3.477463312368973e-05,
      "loss": 0.4222,
      "num_input_tokens_seen": 1743216,
      "step": 2655
    },
    {
      "epoch": 1.3941299790356394,
      "grad_norm": 0.7147731781005859,
      "learning_rate": 3.484014675052411e-05,
      "loss": 0.6232,
      "num_input_tokens_seen": 1746000,
      "step": 2660
    },
    {
      "epoch": 1.3967505241090148,
      "grad_norm": 0.8008643388748169,
      "learning_rate": 3.490566037735849e-05,
      "loss": 0.4775,
      "num_input_tokens_seen": 1748560,
      "step": 2665
    },
    {
      "epoch": 1.39937106918239,
      "grad_norm": 0.4503737688064575,
      "learning_rate": 3.497117400419288e-05,
      "loss": 0.6458,
      "num_input_tokens_seen": 1751504,
      "step": 2670
    },
    {
      "epoch": 1.4019916142557651,
      "grad_norm": 0.8265485763549805,
      "learning_rate": 3.5036687631027254e-05,
      "loss": 0.5168,
      "num_input_tokens_seen": 1754864,
      "step": 2675
    },
    {
      "epoch": 1.4046121593291405,
      "grad_norm": 1.1127448081970215,
      "learning_rate": 3.510220125786164e-05,
      "loss": 0.5944,
      "num_input_tokens_seen": 1757648,
      "step": 2680
    },
    {
      "epoch": 1.4072327044025157,
      "grad_norm": 0.40719228982925415,
      "learning_rate": 3.516771488469602e-05,
      "loss": 0.5462,
      "num_input_tokens_seen": 1760432,
      "step": 2685
    },
    {
      "epoch": 1.409853249475891,
      "grad_norm": 0.3197106719017029,
      "learning_rate": 3.5233228511530397e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 1763664,
      "step": 2690
    },
    {
      "epoch": 1.4124737945492662,
      "grad_norm": 0.39534875750541687,
      "learning_rate": 3.529874213836478e-05,
      "loss": 0.4562,
      "num_input_tokens_seen": 1766640,
      "step": 2695
    },
    {
      "epoch": 1.4150943396226414,
      "grad_norm": 0.5781155824661255,
      "learning_rate": 3.536425576519916e-05,
      "loss": 0.5586,
      "num_input_tokens_seen": 1769776,
      "step": 2700
    },
    {
      "epoch": 1.4177148846960168,
      "grad_norm": 0.7179520130157471,
      "learning_rate": 3.5429769392033546e-05,
      "loss": 0.5764,
      "num_input_tokens_seen": 1772112,
      "step": 2705
    },
    {
      "epoch": 1.420335429769392,
      "grad_norm": 0.4097619950771332,
      "learning_rate": 3.549528301886792e-05,
      "loss": 0.5104,
      "num_input_tokens_seen": 1775312,
      "step": 2710
    },
    {
      "epoch": 1.4229559748427674,
      "grad_norm": 0.46620652079582214,
      "learning_rate": 3.556079664570231e-05,
      "loss": 0.6498,
      "num_input_tokens_seen": 1779504,
      "step": 2715
    },
    {
      "epoch": 1.4255765199161425,
      "grad_norm": 0.4158481955528259,
      "learning_rate": 3.562631027253669e-05,
      "loss": 0.475,
      "num_input_tokens_seen": 1783248,
      "step": 2720
    },
    {
      "epoch": 1.4281970649895177,
      "grad_norm": 0.8309277892112732,
      "learning_rate": 3.569182389937107e-05,
      "loss": 0.3908,
      "num_input_tokens_seen": 1789168,
      "step": 2725
    },
    {
      "epoch": 1.430817610062893,
      "grad_norm": 0.5266585946083069,
      "learning_rate": 3.5757337526205454e-05,
      "loss": 0.5142,
      "num_input_tokens_seen": 1792592,
      "step": 2730
    },
    {
      "epoch": 1.4334381551362683,
      "grad_norm": 1.0133724212646484,
      "learning_rate": 3.582285115303983e-05,
      "loss": 0.5617,
      "num_input_tokens_seen": 1795536,
      "step": 2735
    },
    {
      "epoch": 1.4360587002096437,
      "grad_norm": 0.4545309543609619,
      "learning_rate": 3.588836477987422e-05,
      "loss": 0.4636,
      "num_input_tokens_seen": 1798064,
      "step": 2740
    },
    {
      "epoch": 1.4386792452830188,
      "grad_norm": 0.48522740602493286,
      "learning_rate": 3.59538784067086e-05,
      "loss": 0.5587,
      "num_input_tokens_seen": 1801392,
      "step": 2745
    },
    {
      "epoch": 1.441299790356394,
      "grad_norm": 0.6711395382881165,
      "learning_rate": 3.601939203354298e-05,
      "loss": 0.5703,
      "num_input_tokens_seen": 1804368,
      "step": 2750
    },
    {
      "epoch": 1.4439203354297694,
      "grad_norm": 0.6638684272766113,
      "learning_rate": 3.608490566037736e-05,
      "loss": 0.599,
      "num_input_tokens_seen": 1807504,
      "step": 2755
    },
    {
      "epoch": 1.4465408805031448,
      "grad_norm": 0.5822736024856567,
      "learning_rate": 3.615041928721174e-05,
      "loss": 0.4121,
      "num_input_tokens_seen": 1810160,
      "step": 2760
    },
    {
      "epoch": 1.44916142557652,
      "grad_norm": 0.6072854995727539,
      "learning_rate": 3.621593291404612e-05,
      "loss": 0.5439,
      "num_input_tokens_seen": 1813616,
      "step": 2765
    },
    {
      "epoch": 1.4517819706498951,
      "grad_norm": 0.43810829520225525,
      "learning_rate": 3.6281446540880505e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 1816752,
      "step": 2770
    },
    {
      "epoch": 1.4544025157232705,
      "grad_norm": 0.8228722810745239,
      "learning_rate": 3.634696016771489e-05,
      "loss": 0.5103,
      "num_input_tokens_seen": 1820784,
      "step": 2775
    },
    {
      "epoch": 1.4570230607966457,
      "grad_norm": 0.6385607123374939,
      "learning_rate": 3.6412473794549265e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 1823024,
      "step": 2780
    },
    {
      "epoch": 1.459643605870021,
      "grad_norm": 0.655105710029602,
      "learning_rate": 3.647798742138365e-05,
      "loss": 0.5827,
      "num_input_tokens_seen": 1828016,
      "step": 2785
    },
    {
      "epoch": 1.4622641509433962,
      "grad_norm": 0.35867252945899963,
      "learning_rate": 3.654350104821803e-05,
      "loss": 0.5658,
      "num_input_tokens_seen": 1831280,
      "step": 2790
    },
    {
      "epoch": 1.4648846960167714,
      "grad_norm": 0.3410692512989044,
      "learning_rate": 3.6609014675052414e-05,
      "loss": 0.4533,
      "num_input_tokens_seen": 1834896,
      "step": 2795
    },
    {
      "epoch": 1.4675052410901468,
      "grad_norm": 0.47645103931427,
      "learning_rate": 3.66745283018868e-05,
      "loss": 0.5376,
      "num_input_tokens_seen": 1837488,
      "step": 2800
    },
    {
      "epoch": 1.470125786163522,
      "grad_norm": 0.45423582196235657,
      "learning_rate": 3.6740041928721173e-05,
      "loss": 0.5971,
      "num_input_tokens_seen": 1841584,
      "step": 2805
    },
    {
      "epoch": 1.4727463312368974,
      "grad_norm": 0.600600004196167,
      "learning_rate": 3.6805555555555556e-05,
      "loss": 0.593,
      "num_input_tokens_seen": 1844688,
      "step": 2810
    },
    {
      "epoch": 1.4753668763102725,
      "grad_norm": 0.45546936988830566,
      "learning_rate": 3.687106918238994e-05,
      "loss": 0.5379,
      "num_input_tokens_seen": 1847472,
      "step": 2815
    },
    {
      "epoch": 1.4779874213836477,
      "grad_norm": 0.6941242814064026,
      "learning_rate": 3.6936582809224316e-05,
      "loss": 0.7946,
      "num_input_tokens_seen": 1851088,
      "step": 2820
    },
    {
      "epoch": 1.480607966457023,
      "grad_norm": 0.4668252170085907,
      "learning_rate": 3.7002096436058706e-05,
      "loss": 0.5415,
      "num_input_tokens_seen": 1855216,
      "step": 2825
    },
    {
      "epoch": 1.4832285115303983,
      "grad_norm": 0.4331161379814148,
      "learning_rate": 3.706761006289308e-05,
      "loss": 0.4672,
      "num_input_tokens_seen": 1857744,
      "step": 2830
    },
    {
      "epoch": 1.4858490566037736,
      "grad_norm": 0.4974406063556671,
      "learning_rate": 3.7133123689727465e-05,
      "loss": 0.4879,
      "num_input_tokens_seen": 1860400,
      "step": 2835
    },
    {
      "epoch": 1.4884696016771488,
      "grad_norm": 0.5817961096763611,
      "learning_rate": 3.719863731656185e-05,
      "loss": 0.5822,
      "num_input_tokens_seen": 1863248,
      "step": 2840
    },
    {
      "epoch": 1.491090146750524,
      "grad_norm": 0.44133782386779785,
      "learning_rate": 3.7264150943396224e-05,
      "loss": 0.5273,
      "num_input_tokens_seen": 1866576,
      "step": 2845
    },
    {
      "epoch": 1.4937106918238994,
      "grad_norm": 0.476677268743515,
      "learning_rate": 3.732966457023061e-05,
      "loss": 0.4972,
      "num_input_tokens_seen": 1869840,
      "step": 2850
    },
    {
      "epoch": 1.4963312368972748,
      "grad_norm": 0.41519564390182495,
      "learning_rate": 3.739517819706499e-05,
      "loss": 0.6541,
      "num_input_tokens_seen": 1872624,
      "step": 2855
    },
    {
      "epoch": 1.49895178197065,
      "grad_norm": 0.35330235958099365,
      "learning_rate": 3.7460691823899374e-05,
      "loss": 0.5501,
      "num_input_tokens_seen": 1876016,
      "step": 2860
    },
    {
      "epoch": 1.501572327044025,
      "grad_norm": 0.43081146478652954,
      "learning_rate": 3.752620545073376e-05,
      "loss": 0.4157,
      "num_input_tokens_seen": 1880976,
      "step": 2865
    },
    {
      "epoch": 1.5041928721174003,
      "grad_norm": 1.2389250993728638,
      "learning_rate": 3.759171907756814e-05,
      "loss": 0.6227,
      "num_input_tokens_seen": 1883568,
      "step": 2870
    },
    {
      "epoch": 1.5068134171907757,
      "grad_norm": 0.7389305830001831,
      "learning_rate": 3.7657232704402516e-05,
      "loss": 0.5094,
      "num_input_tokens_seen": 1886800,
      "step": 2875
    },
    {
      "epoch": 1.509433962264151,
      "grad_norm": 0.6096968650817871,
      "learning_rate": 3.77227463312369e-05,
      "loss": 0.5161,
      "num_input_tokens_seen": 1890672,
      "step": 2880
    },
    {
      "epoch": 1.5120545073375262,
      "grad_norm": 0.6269100308418274,
      "learning_rate": 3.778825995807128e-05,
      "loss": 0.3976,
      "num_input_tokens_seen": 1893456,
      "step": 2885
    },
    {
      "epoch": 1.5146750524109014,
      "grad_norm": 0.5686066746711731,
      "learning_rate": 3.785377358490566e-05,
      "loss": 0.5532,
      "num_input_tokens_seen": 1896752,
      "step": 2890
    },
    {
      "epoch": 1.5172955974842768,
      "grad_norm": 0.28543707728385925,
      "learning_rate": 3.791928721174005e-05,
      "loss": 0.4659,
      "num_input_tokens_seen": 1899664,
      "step": 2895
    },
    {
      "epoch": 1.519916142557652,
      "grad_norm": 0.3654082119464874,
      "learning_rate": 3.7984800838574425e-05,
      "loss": 0.4312,
      "num_input_tokens_seen": 1902320,
      "step": 2900
    },
    {
      "epoch": 1.5225366876310273,
      "grad_norm": 0.3661927580833435,
      "learning_rate": 3.805031446540881e-05,
      "loss": 0.5797,
      "num_input_tokens_seen": 1905488,
      "step": 2905
    },
    {
      "epoch": 1.5251572327044025,
      "grad_norm": 0.4208669364452362,
      "learning_rate": 3.811582809224319e-05,
      "loss": 0.5114,
      "num_input_tokens_seen": 1908144,
      "step": 2910
    },
    {
      "epoch": 1.5277777777777777,
      "grad_norm": 0.32450374960899353,
      "learning_rate": 3.818134171907757e-05,
      "loss": 0.5297,
      "num_input_tokens_seen": 1911536,
      "step": 2915
    },
    {
      "epoch": 1.530398322851153,
      "grad_norm": 0.3136166036128998,
      "learning_rate": 3.824685534591195e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 1916176,
      "step": 2920
    },
    {
      "epoch": 1.5330188679245285,
      "grad_norm": 0.46879956126213074,
      "learning_rate": 3.831236897274633e-05,
      "loss": 0.5224,
      "num_input_tokens_seen": 1920592,
      "step": 2925
    },
    {
      "epoch": 1.5356394129979036,
      "grad_norm": 0.5701702833175659,
      "learning_rate": 3.8377882599580716e-05,
      "loss": 0.4853,
      "num_input_tokens_seen": 1923152,
      "step": 2930
    },
    {
      "epoch": 1.5382599580712788,
      "grad_norm": 0.6799443364143372,
      "learning_rate": 3.844339622641509e-05,
      "loss": 0.563,
      "num_input_tokens_seen": 1926000,
      "step": 2935
    },
    {
      "epoch": 1.540880503144654,
      "grad_norm": 0.45838049054145813,
      "learning_rate": 3.8508909853249476e-05,
      "loss": 0.6173,
      "num_input_tokens_seen": 1929072,
      "step": 2940
    },
    {
      "epoch": 1.5435010482180294,
      "grad_norm": 0.26664048433303833,
      "learning_rate": 3.857442348008386e-05,
      "loss": 0.396,
      "num_input_tokens_seen": 1932208,
      "step": 2945
    },
    {
      "epoch": 1.5461215932914047,
      "grad_norm": 0.45398861169815063,
      "learning_rate": 3.863993710691824e-05,
      "loss": 0.6932,
      "num_input_tokens_seen": 1935152,
      "step": 2950
    },
    {
      "epoch": 1.54874213836478,
      "grad_norm": 0.576120913028717,
      "learning_rate": 3.8705450733752625e-05,
      "loss": 0.5684,
      "num_input_tokens_seen": 1938704,
      "step": 2955
    },
    {
      "epoch": 1.551362683438155,
      "grad_norm": 0.3741660714149475,
      "learning_rate": 3.8770964360587e-05,
      "loss": 0.6207,
      "num_input_tokens_seen": 1942192,
      "step": 2960
    },
    {
      "epoch": 1.5539832285115303,
      "grad_norm": 0.8084535598754883,
      "learning_rate": 3.883647798742139e-05,
      "loss": 0.4834,
      "num_input_tokens_seen": 1945744,
      "step": 2965
    },
    {
      "epoch": 1.5566037735849056,
      "grad_norm": 0.4400956630706787,
      "learning_rate": 3.890199161425577e-05,
      "loss": 0.5204,
      "num_input_tokens_seen": 1948944,
      "step": 2970
    },
    {
      "epoch": 1.559224318658281,
      "grad_norm": 0.5314616560935974,
      "learning_rate": 3.8967505241090144e-05,
      "loss": 0.4872,
      "num_input_tokens_seen": 1951792,
      "step": 2975
    },
    {
      "epoch": 1.5618448637316562,
      "grad_norm": 0.5116677284240723,
      "learning_rate": 3.9033018867924534e-05,
      "loss": 0.5949,
      "num_input_tokens_seen": 1955376,
      "step": 2980
    },
    {
      "epoch": 1.5644654088050314,
      "grad_norm": 0.432243674993515,
      "learning_rate": 3.909853249475891e-05,
      "loss": 0.5459,
      "num_input_tokens_seen": 1959120,
      "step": 2985
    },
    {
      "epoch": 1.5670859538784065,
      "grad_norm": 0.2655541002750397,
      "learning_rate": 3.916404612159329e-05,
      "loss": 0.423,
      "num_input_tokens_seen": 1964304,
      "step": 2990
    },
    {
      "epoch": 1.569706498951782,
      "grad_norm": 0.5467645525932312,
      "learning_rate": 3.9229559748427676e-05,
      "loss": 0.6711,
      "num_input_tokens_seen": 1966832,
      "step": 2995
    },
    {
      "epoch": 1.5723270440251573,
      "grad_norm": 0.4997488558292389,
      "learning_rate": 3.929507337526205e-05,
      "loss": 0.6271,
      "num_input_tokens_seen": 1970224,
      "step": 3000
    },
    {
      "epoch": 1.5749475890985325,
      "grad_norm": 0.29031023383140564,
      "learning_rate": 3.9360587002096435e-05,
      "loss": 0.5471,
      "num_input_tokens_seen": 1974288,
      "step": 3005
    },
    {
      "epoch": 1.5775681341719077,
      "grad_norm": 0.5472842454910278,
      "learning_rate": 3.942610062893082e-05,
      "loss": 0.4643,
      "num_input_tokens_seen": 1978064,
      "step": 3010
    },
    {
      "epoch": 1.580188679245283,
      "grad_norm": 0.5420131683349609,
      "learning_rate": 3.94916142557652e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 1981744,
      "step": 3015
    },
    {
      "epoch": 1.5828092243186582,
      "grad_norm": 0.9199143648147583,
      "learning_rate": 3.9557127882599585e-05,
      "loss": 0.5722,
      "num_input_tokens_seen": 1984016,
      "step": 3020
    },
    {
      "epoch": 1.5854297693920336,
      "grad_norm": 0.576610267162323,
      "learning_rate": 3.962264150943397e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 1987056,
      "step": 3025
    },
    {
      "epoch": 1.5880503144654088,
      "grad_norm": 0.39660120010375977,
      "learning_rate": 3.9688155136268344e-05,
      "loss": 0.5655,
      "num_input_tokens_seen": 1992048,
      "step": 3030
    },
    {
      "epoch": 1.590670859538784,
      "grad_norm": 0.3556070625782013,
      "learning_rate": 3.975366876310273e-05,
      "loss": 0.7015,
      "num_input_tokens_seen": 1995472,
      "step": 3035
    },
    {
      "epoch": 1.5932914046121593,
      "grad_norm": 0.3714050054550171,
      "learning_rate": 3.981918238993711e-05,
      "loss": 0.4068,
      "num_input_tokens_seen": 1999152,
      "step": 3040
    },
    {
      "epoch": 1.5959119496855347,
      "grad_norm": 0.5103093981742859,
      "learning_rate": 3.9884696016771486e-05,
      "loss": 0.5914,
      "num_input_tokens_seen": 2001744,
      "step": 3045
    },
    {
      "epoch": 1.59853249475891,
      "grad_norm": 0.4481639266014099,
      "learning_rate": 3.9950209643605876e-05,
      "loss": 0.6412,
      "num_input_tokens_seen": 2004848,
      "step": 3050
    },
    {
      "epoch": 1.601153039832285,
      "grad_norm": 0.7659627795219421,
      "learning_rate": 4.001572327044025e-05,
      "loss": 0.5635,
      "num_input_tokens_seen": 2007920,
      "step": 3055
    },
    {
      "epoch": 1.6037735849056602,
      "grad_norm": 0.4613637924194336,
      "learning_rate": 4.0081236897274636e-05,
      "loss": 0.5057,
      "num_input_tokens_seen": 2011152,
      "step": 3060
    },
    {
      "epoch": 1.6063941299790356,
      "grad_norm": 0.6514269113540649,
      "learning_rate": 4.014675052410902e-05,
      "loss": 0.6269,
      "num_input_tokens_seen": 2013968,
      "step": 3065
    },
    {
      "epoch": 1.609014675052411,
      "grad_norm": 0.7040261626243591,
      "learning_rate": 4.0212264150943395e-05,
      "loss": 0.5491,
      "num_input_tokens_seen": 2017936,
      "step": 3070
    },
    {
      "epoch": 1.6116352201257862,
      "grad_norm": 0.2714079022407532,
      "learning_rate": 4.027777777777778e-05,
      "loss": 0.5355,
      "num_input_tokens_seen": 2020528,
      "step": 3075
    },
    {
      "epoch": 1.6142557651991614,
      "grad_norm": 0.617059051990509,
      "learning_rate": 4.034329140461216e-05,
      "loss": 0.6914,
      "num_input_tokens_seen": 2023792,
      "step": 3080
    },
    {
      "epoch": 1.6168763102725365,
      "grad_norm": 0.44810041785240173,
      "learning_rate": 4.0408805031446544e-05,
      "loss": 0.5971,
      "num_input_tokens_seen": 2027152,
      "step": 3085
    },
    {
      "epoch": 1.619496855345912,
      "grad_norm": 0.34634461998939514,
      "learning_rate": 4.047431865828093e-05,
      "loss": 0.4484,
      "num_input_tokens_seen": 2029968,
      "step": 3090
    },
    {
      "epoch": 1.6221174004192873,
      "grad_norm": 0.6549427509307861,
      "learning_rate": 4.0539832285115304e-05,
      "loss": 0.4873,
      "num_input_tokens_seen": 2033488,
      "step": 3095
    },
    {
      "epoch": 1.6247379454926625,
      "grad_norm": 0.9860314726829529,
      "learning_rate": 4.0605345911949687e-05,
      "loss": 0.63,
      "num_input_tokens_seen": 2037136,
      "step": 3100
    },
    {
      "epoch": 1.6273584905660377,
      "grad_norm": 0.41516029834747314,
      "learning_rate": 4.067085953878407e-05,
      "loss": 0.5449,
      "num_input_tokens_seen": 2039888,
      "step": 3105
    },
    {
      "epoch": 1.629979035639413,
      "grad_norm": 0.42100852727890015,
      "learning_rate": 4.073637316561845e-05,
      "loss": 0.7115,
      "num_input_tokens_seen": 2043024,
      "step": 3110
    },
    {
      "epoch": 1.6325995807127882,
      "grad_norm": 0.4484747648239136,
      "learning_rate": 4.080188679245283e-05,
      "loss": 0.4314,
      "num_input_tokens_seen": 2047056,
      "step": 3115
    },
    {
      "epoch": 1.6352201257861636,
      "grad_norm": 1.012222409248352,
      "learning_rate": 4.086740041928722e-05,
      "loss": 0.5904,
      "num_input_tokens_seen": 2050000,
      "step": 3120
    },
    {
      "epoch": 1.6378406708595388,
      "grad_norm": 0.4013672471046448,
      "learning_rate": 4.0932914046121595e-05,
      "loss": 0.5298,
      "num_input_tokens_seen": 2054032,
      "step": 3125
    },
    {
      "epoch": 1.640461215932914,
      "grad_norm": 0.49525249004364014,
      "learning_rate": 4.099842767295597e-05,
      "loss": 0.526,
      "num_input_tokens_seen": 2057072,
      "step": 3130
    },
    {
      "epoch": 1.6430817610062893,
      "grad_norm": 0.33458974957466125,
      "learning_rate": 4.106394129979036e-05,
      "loss": 0.4707,
      "num_input_tokens_seen": 2060656,
      "step": 3135
    },
    {
      "epoch": 1.6457023060796647,
      "grad_norm": 1.1402965784072876,
      "learning_rate": 4.112945492662474e-05,
      "loss": 0.541,
      "num_input_tokens_seen": 2063568,
      "step": 3140
    },
    {
      "epoch": 1.64832285115304,
      "grad_norm": 0.49002543091773987,
      "learning_rate": 4.119496855345912e-05,
      "loss": 0.4518,
      "num_input_tokens_seen": 2066960,
      "step": 3145
    },
    {
      "epoch": 1.650943396226415,
      "grad_norm": 0.4477578103542328,
      "learning_rate": 4.1260482180293504e-05,
      "loss": 0.5244,
      "num_input_tokens_seen": 2070096,
      "step": 3150
    },
    {
      "epoch": 1.6535639412997902,
      "grad_norm": 0.3953956663608551,
      "learning_rate": 4.132599580712788e-05,
      "loss": 0.5255,
      "num_input_tokens_seen": 2073072,
      "step": 3155
    },
    {
      "epoch": 1.6561844863731656,
      "grad_norm": 0.4120720326900482,
      "learning_rate": 4.139150943396227e-05,
      "loss": 0.478,
      "num_input_tokens_seen": 2075632,
      "step": 3160
    },
    {
      "epoch": 1.658805031446541,
      "grad_norm": 0.7818681001663208,
      "learning_rate": 4.1457023060796646e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 2079024,
      "step": 3165
    },
    {
      "epoch": 1.6614255765199162,
      "grad_norm": 0.782138466835022,
      "learning_rate": 4.152253668763103e-05,
      "loss": 0.6815,
      "num_input_tokens_seen": 2081744,
      "step": 3170
    },
    {
      "epoch": 1.6640461215932913,
      "grad_norm": 0.3284837305545807,
      "learning_rate": 4.158805031446541e-05,
      "loss": 0.5797,
      "num_input_tokens_seen": 2084784,
      "step": 3175
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 0.36621949076652527,
      "learning_rate": 4.1653563941299795e-05,
      "loss": 0.4626,
      "num_input_tokens_seen": 2087984,
      "step": 3180
    },
    {
      "epoch": 1.669287211740042,
      "grad_norm": 0.33338120579719543,
      "learning_rate": 4.171907756813417e-05,
      "loss": 0.5279,
      "num_input_tokens_seen": 2091344,
      "step": 3185
    },
    {
      "epoch": 1.6719077568134173,
      "grad_norm": 0.32382676005363464,
      "learning_rate": 4.1784591194968555e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 2095312,
      "step": 3190
    },
    {
      "epoch": 1.6745283018867925,
      "grad_norm": 0.3253656327724457,
      "learning_rate": 4.185010482180294e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 2098832,
      "step": 3195
    },
    {
      "epoch": 1.6771488469601676,
      "grad_norm": 0.6694507598876953,
      "learning_rate": 4.1915618448637314e-05,
      "loss": 0.6789,
      "num_input_tokens_seen": 2101168,
      "step": 3200
    },
    {
      "epoch": 1.679769392033543,
      "grad_norm": 0.6728124618530273,
      "learning_rate": 4.1981132075471704e-05,
      "loss": 0.7636,
      "num_input_tokens_seen": 2104816,
      "step": 3205
    },
    {
      "epoch": 1.6823899371069182,
      "grad_norm": 0.7681397199630737,
      "learning_rate": 4.204664570230608e-05,
      "loss": 0.524,
      "num_input_tokens_seen": 2107728,
      "step": 3210
    },
    {
      "epoch": 1.6850104821802936,
      "grad_norm": 1.1711385250091553,
      "learning_rate": 4.2112159329140463e-05,
      "loss": 0.773,
      "num_input_tokens_seen": 2110992,
      "step": 3215
    },
    {
      "epoch": 1.6876310272536688,
      "grad_norm": 0.6837564706802368,
      "learning_rate": 4.2177672955974846e-05,
      "loss": 0.5156,
      "num_input_tokens_seen": 2113840,
      "step": 3220
    },
    {
      "epoch": 1.690251572327044,
      "grad_norm": 0.5433861017227173,
      "learning_rate": 4.224318658280922e-05,
      "loss": 0.5163,
      "num_input_tokens_seen": 2116784,
      "step": 3225
    },
    {
      "epoch": 1.6928721174004193,
      "grad_norm": 0.3038249909877777,
      "learning_rate": 4.2308700209643606e-05,
      "loss": 0.6616,
      "num_input_tokens_seen": 2119696,
      "step": 3230
    },
    {
      "epoch": 1.6954926624737947,
      "grad_norm": 0.5100158452987671,
      "learning_rate": 4.237421383647799e-05,
      "loss": 0.5762,
      "num_input_tokens_seen": 2122608,
      "step": 3235
    },
    {
      "epoch": 1.6981132075471699,
      "grad_norm": 0.49778512120246887,
      "learning_rate": 4.243972746331237e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 2125808,
      "step": 3240
    },
    {
      "epoch": 1.700733752620545,
      "grad_norm": 0.593447744846344,
      "learning_rate": 4.2505241090146755e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 2129584,
      "step": 3245
    },
    {
      "epoch": 1.7033542976939202,
      "grad_norm": 0.49167394638061523,
      "learning_rate": 4.257075471698113e-05,
      "loss": 0.5047,
      "num_input_tokens_seen": 2132208,
      "step": 3250
    },
    {
      "epoch": 1.7059748427672956,
      "grad_norm": 0.3355075418949127,
      "learning_rate": 4.2636268343815514e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 2135632,
      "step": 3255
    },
    {
      "epoch": 1.708595387840671,
      "grad_norm": 0.6322067379951477,
      "learning_rate": 4.27017819706499e-05,
      "loss": 0.7281,
      "num_input_tokens_seen": 2139760,
      "step": 3260
    },
    {
      "epoch": 1.7112159329140462,
      "grad_norm": 0.5544403791427612,
      "learning_rate": 4.276729559748428e-05,
      "loss": 0.5233,
      "num_input_tokens_seen": 2143280,
      "step": 3265
    },
    {
      "epoch": 1.7138364779874213,
      "grad_norm": 0.44479793310165405,
      "learning_rate": 4.283280922431866e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 2145648,
      "step": 3270
    },
    {
      "epoch": 1.7164570230607965,
      "grad_norm": 0.642645537853241,
      "learning_rate": 4.289832285115305e-05,
      "loss": 0.5136,
      "num_input_tokens_seen": 2149040,
      "step": 3275
    },
    {
      "epoch": 1.719077568134172,
      "grad_norm": 0.35382261872291565,
      "learning_rate": 4.296383647798742e-05,
      "loss": 0.6149,
      "num_input_tokens_seen": 2151792,
      "step": 3280
    },
    {
      "epoch": 1.7216981132075473,
      "grad_norm": 0.6413965225219727,
      "learning_rate": 4.3029350104821806e-05,
      "loss": 0.4797,
      "num_input_tokens_seen": 2155600,
      "step": 3285
    },
    {
      "epoch": 1.7243186582809225,
      "grad_norm": 0.3873487710952759,
      "learning_rate": 4.309486373165619e-05,
      "loss": 0.4448,
      "num_input_tokens_seen": 2159088,
      "step": 3290
    },
    {
      "epoch": 1.7269392033542976,
      "grad_norm": 0.6197674870491028,
      "learning_rate": 4.3160377358490565e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 2162800,
      "step": 3295
    },
    {
      "epoch": 1.7295597484276728,
      "grad_norm": 0.5707959532737732,
      "learning_rate": 4.322589098532495e-05,
      "loss": 0.5798,
      "num_input_tokens_seen": 2165872,
      "step": 3300
    },
    {
      "epoch": 1.7321802935010482,
      "grad_norm": 0.4667976200580597,
      "learning_rate": 4.329140461215933e-05,
      "loss": 0.6191,
      "num_input_tokens_seen": 2169456,
      "step": 3305
    },
    {
      "epoch": 1.7348008385744236,
      "grad_norm": 0.3716704845428467,
      "learning_rate": 4.335691823899371e-05,
      "loss": 0.5092,
      "num_input_tokens_seen": 2172368,
      "step": 3310
    },
    {
      "epoch": 1.7374213836477987,
      "grad_norm": 0.32113680243492126,
      "learning_rate": 4.34224318658281e-05,
      "loss": 0.6358,
      "num_input_tokens_seen": 2175504,
      "step": 3315
    },
    {
      "epoch": 1.740041928721174,
      "grad_norm": 0.6977897882461548,
      "learning_rate": 4.3487945492662474e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 2179440,
      "step": 3320
    },
    {
      "epoch": 1.7426624737945493,
      "grad_norm": 0.6785503029823303,
      "learning_rate": 4.355345911949686e-05,
      "loss": 0.636,
      "num_input_tokens_seen": 2181904,
      "step": 3325
    },
    {
      "epoch": 1.7452830188679245,
      "grad_norm": 0.4147198796272278,
      "learning_rate": 4.361897274633124e-05,
      "loss": 0.4329,
      "num_input_tokens_seen": 2184816,
      "step": 3330
    },
    {
      "epoch": 1.7479035639412999,
      "grad_norm": 0.9568979740142822,
      "learning_rate": 4.368448637316562e-05,
      "loss": 0.6224,
      "num_input_tokens_seen": 2187920,
      "step": 3335
    },
    {
      "epoch": 1.750524109014675,
      "grad_norm": 0.2383287250995636,
      "learning_rate": 4.375e-05,
      "loss": 0.5476,
      "num_input_tokens_seen": 2191216,
      "step": 3340
    },
    {
      "epoch": 1.7531446540880502,
      "grad_norm": 0.5442902445793152,
      "learning_rate": 4.381551362683438e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 2193744,
      "step": 3345
    },
    {
      "epoch": 1.7557651991614256,
      "grad_norm": 0.5598207116127014,
      "learning_rate": 4.3881027253668766e-05,
      "loss": 0.4874,
      "num_input_tokens_seen": 2196368,
      "step": 3350
    },
    {
      "epoch": 1.758385744234801,
      "grad_norm": 0.3705276548862457,
      "learning_rate": 4.394654088050314e-05,
      "loss": 0.3943,
      "num_input_tokens_seen": 2199632,
      "step": 3355
    },
    {
      "epoch": 1.7610062893081762,
      "grad_norm": 0.35693418979644775,
      "learning_rate": 4.401205450733753e-05,
      "loss": 0.5319,
      "num_input_tokens_seen": 2202704,
      "step": 3360
    },
    {
      "epoch": 1.7636268343815513,
      "grad_norm": 0.48763972520828247,
      "learning_rate": 4.407756813417191e-05,
      "loss": 0.7184,
      "num_input_tokens_seen": 2205616,
      "step": 3365
    },
    {
      "epoch": 1.7662473794549265,
      "grad_norm": 0.32889625430107117,
      "learning_rate": 4.414308176100629e-05,
      "loss": 0.4854,
      "num_input_tokens_seen": 2209008,
      "step": 3370
    },
    {
      "epoch": 1.7688679245283019,
      "grad_norm": 0.4335348904132843,
      "learning_rate": 4.4208595387840674e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 2211824,
      "step": 3375
    },
    {
      "epoch": 1.7714884696016773,
      "grad_norm": 0.3395770788192749,
      "learning_rate": 4.427410901467505e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 2215632,
      "step": 3380
    },
    {
      "epoch": 1.7741090146750524,
      "grad_norm": 0.5756599307060242,
      "learning_rate": 4.433962264150944e-05,
      "loss": 0.4827,
      "num_input_tokens_seen": 2218512,
      "step": 3385
    },
    {
      "epoch": 1.7767295597484276,
      "grad_norm": 0.37732353806495667,
      "learning_rate": 4.440513626834382e-05,
      "loss": 0.4245,
      "num_input_tokens_seen": 2221776,
      "step": 3390
    },
    {
      "epoch": 1.7793501048218028,
      "grad_norm": 0.2688818871974945,
      "learning_rate": 4.44706498951782e-05,
      "loss": 0.4563,
      "num_input_tokens_seen": 2224784,
      "step": 3395
    },
    {
      "epoch": 1.7819706498951782,
      "grad_norm": 0.8615556955337524,
      "learning_rate": 4.453616352201258e-05,
      "loss": 0.5251,
      "num_input_tokens_seen": 2227408,
      "step": 3400
    },
    {
      "epoch": 1.7845911949685536,
      "grad_norm": 0.6367427706718445,
      "learning_rate": 4.460167714884696e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 2230064,
      "step": 3405
    },
    {
      "epoch": 1.7872117400419287,
      "grad_norm": 0.5759003162384033,
      "learning_rate": 4.466719077568134e-05,
      "loss": 0.4123,
      "num_input_tokens_seen": 2234288,
      "step": 3410
    },
    {
      "epoch": 1.789832285115304,
      "grad_norm": 0.3201746940612793,
      "learning_rate": 4.4732704402515725e-05,
      "loss": 0.4294,
      "num_input_tokens_seen": 2237072,
      "step": 3415
    },
    {
      "epoch": 1.7924528301886793,
      "grad_norm": 0.33222341537475586,
      "learning_rate": 4.479821802935011e-05,
      "loss": 0.406,
      "num_input_tokens_seen": 2240016,
      "step": 3420
    },
    {
      "epoch": 1.7950733752620545,
      "grad_norm": 0.42930877208709717,
      "learning_rate": 4.4863731656184485e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 2242800,
      "step": 3425
    },
    {
      "epoch": 1.7976939203354299,
      "grad_norm": 0.7239888906478882,
      "learning_rate": 4.4929245283018875e-05,
      "loss": 0.5803,
      "num_input_tokens_seen": 2245680,
      "step": 3430
    },
    {
      "epoch": 1.800314465408805,
      "grad_norm": 0.6114919781684875,
      "learning_rate": 4.499475890985325e-05,
      "loss": 0.5452,
      "num_input_tokens_seen": 2248272,
      "step": 3435
    },
    {
      "epoch": 1.8029350104821802,
      "grad_norm": 0.8098805546760559,
      "learning_rate": 4.5060272536687634e-05,
      "loss": 0.3733,
      "num_input_tokens_seen": 2251376,
      "step": 3440
    },
    {
      "epoch": 1.8055555555555556,
      "grad_norm": 0.7120617032051086,
      "learning_rate": 4.512578616352202e-05,
      "loss": 0.5775,
      "num_input_tokens_seen": 2254256,
      "step": 3445
    },
    {
      "epoch": 1.808176100628931,
      "grad_norm": 0.6881124973297119,
      "learning_rate": 4.519129979035639e-05,
      "loss": 0.5976,
      "num_input_tokens_seen": 2257104,
      "step": 3450
    },
    {
      "epoch": 1.8107966457023061,
      "grad_norm": 0.4672967195510864,
      "learning_rate": 4.525681341719078e-05,
      "loss": 0.5856,
      "num_input_tokens_seen": 2259920,
      "step": 3455
    },
    {
      "epoch": 1.8134171907756813,
      "grad_norm": 0.34158921241760254,
      "learning_rate": 4.532232704402516e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 2262736,
      "step": 3460
    },
    {
      "epoch": 1.8160377358490565,
      "grad_norm": 0.23368576169013977,
      "learning_rate": 4.5387840670859536e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 2266864,
      "step": 3465
    },
    {
      "epoch": 1.8186582809224319,
      "grad_norm": 1.0187751054763794,
      "learning_rate": 4.5453354297693926e-05,
      "loss": 0.7202,
      "num_input_tokens_seen": 2269584,
      "step": 3470
    },
    {
      "epoch": 1.8212788259958073,
      "grad_norm": 0.2830042541027069,
      "learning_rate": 4.55188679245283e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 2272720,
      "step": 3475
    },
    {
      "epoch": 1.8238993710691824,
      "grad_norm": 0.34767064452171326,
      "learning_rate": 4.5584381551362685e-05,
      "loss": 0.6113,
      "num_input_tokens_seen": 2276880,
      "step": 3480
    },
    {
      "epoch": 1.8265199161425576,
      "grad_norm": 1.3516294956207275,
      "learning_rate": 4.564989517819707e-05,
      "loss": 0.4374,
      "num_input_tokens_seen": 2280016,
      "step": 3485
    },
    {
      "epoch": 1.8291404612159328,
      "grad_norm": 0.4549809396266937,
      "learning_rate": 4.571540880503145e-05,
      "loss": 0.4749,
      "num_input_tokens_seen": 2282640,
      "step": 3490
    },
    {
      "epoch": 1.8317610062893082,
      "grad_norm": 0.37545642256736755,
      "learning_rate": 4.578092243186583e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 2286032,
      "step": 3495
    },
    {
      "epoch": 1.8343815513626835,
      "grad_norm": 0.45055606961250305,
      "learning_rate": 4.584643605870021e-05,
      "loss": 0.6016,
      "num_input_tokens_seen": 2289136,
      "step": 3500
    },
    {
      "epoch": 1.8370020964360587,
      "grad_norm": 0.6170727014541626,
      "learning_rate": 4.5911949685534594e-05,
      "loss": 0.4157,
      "num_input_tokens_seen": 2292944,
      "step": 3505
    },
    {
      "epoch": 1.8396226415094339,
      "grad_norm": 0.9442532062530518,
      "learning_rate": 4.5977463312368977e-05,
      "loss": 0.4992,
      "num_input_tokens_seen": 2296240,
      "step": 3510
    },
    {
      "epoch": 1.8422431865828093,
      "grad_norm": 1.4527819156646729,
      "learning_rate": 4.604297693920336e-05,
      "loss": 0.5001,
      "num_input_tokens_seen": 2299088,
      "step": 3515
    },
    {
      "epoch": 1.8448637316561844,
      "grad_norm": 0.5698071122169495,
      "learning_rate": 4.6108490566037736e-05,
      "loss": 0.5227,
      "num_input_tokens_seen": 2302736,
      "step": 3520
    },
    {
      "epoch": 1.8474842767295598,
      "grad_norm": 0.4473409056663513,
      "learning_rate": 4.617400419287212e-05,
      "loss": 0.5937,
      "num_input_tokens_seen": 2306640,
      "step": 3525
    },
    {
      "epoch": 1.850104821802935,
      "grad_norm": 0.506717324256897,
      "learning_rate": 4.62395178197065e-05,
      "loss": 0.5726,
      "num_input_tokens_seen": 2309840,
      "step": 3530
    },
    {
      "epoch": 1.8527253668763102,
      "grad_norm": 0.8270758390426636,
      "learning_rate": 4.630503144654088e-05,
      "loss": 0.5591,
      "num_input_tokens_seen": 2312208,
      "step": 3535
    },
    {
      "epoch": 1.8553459119496856,
      "grad_norm": 1.005803108215332,
      "learning_rate": 4.637054507337527e-05,
      "loss": 0.6647,
      "num_input_tokens_seen": 2315088,
      "step": 3540
    },
    {
      "epoch": 1.857966457023061,
      "grad_norm": 0.694409191608429,
      "learning_rate": 4.6436058700209645e-05,
      "loss": 0.5413,
      "num_input_tokens_seen": 2318224,
      "step": 3545
    },
    {
      "epoch": 1.8605870020964361,
      "grad_norm": 0.26028257608413696,
      "learning_rate": 4.650157232704403e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 2322800,
      "step": 3550
    },
    {
      "epoch": 1.8632075471698113,
      "grad_norm": 0.5377190709114075,
      "learning_rate": 4.656708595387841e-05,
      "loss": 0.3539,
      "num_input_tokens_seen": 2329008,
      "step": 3555
    },
    {
      "epoch": 1.8658280922431865,
      "grad_norm": 0.7291828393936157,
      "learning_rate": 4.663259958071279e-05,
      "loss": 0.6672,
      "num_input_tokens_seen": 2331344,
      "step": 3560
    },
    {
      "epoch": 1.8684486373165619,
      "grad_norm": 0.3988966941833496,
      "learning_rate": 4.669811320754717e-05,
      "loss": 0.4983,
      "num_input_tokens_seen": 2334352,
      "step": 3565
    },
    {
      "epoch": 1.8710691823899372,
      "grad_norm": 0.2820865511894226,
      "learning_rate": 4.676362683438155e-05,
      "loss": 0.5926,
      "num_input_tokens_seen": 2338512,
      "step": 3570
    },
    {
      "epoch": 1.8736897274633124,
      "grad_norm": 0.40671229362487793,
      "learning_rate": 4.6829140461215936e-05,
      "loss": 0.6014,
      "num_input_tokens_seen": 2342224,
      "step": 3575
    },
    {
      "epoch": 1.8763102725366876,
      "grad_norm": 0.9812504053115845,
      "learning_rate": 4.689465408805032e-05,
      "loss": 0.4887,
      "num_input_tokens_seen": 2345808,
      "step": 3580
    },
    {
      "epoch": 1.8789308176100628,
      "grad_norm": 0.295613557100296,
      "learning_rate": 4.69601677148847e-05,
      "loss": 0.5779,
      "num_input_tokens_seen": 2348560,
      "step": 3585
    },
    {
      "epoch": 1.8815513626834381,
      "grad_norm": 0.433596134185791,
      "learning_rate": 4.702568134171908e-05,
      "loss": 0.4075,
      "num_input_tokens_seen": 2351824,
      "step": 3590
    },
    {
      "epoch": 1.8841719077568135,
      "grad_norm": 0.8665644526481628,
      "learning_rate": 4.709119496855346e-05,
      "loss": 0.4024,
      "num_input_tokens_seen": 2355696,
      "step": 3595
    },
    {
      "epoch": 1.8867924528301887,
      "grad_norm": 0.48128706216812134,
      "learning_rate": 4.7156708595387845e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 2358352,
      "step": 3600
    },
    {
      "epoch": 1.8894129979035639,
      "grad_norm": 0.33137696981430054,
      "learning_rate": 4.722222222222222e-05,
      "loss": 0.4445,
      "num_input_tokens_seen": 2361776,
      "step": 3605
    },
    {
      "epoch": 1.892033542976939,
      "grad_norm": 0.5357550382614136,
      "learning_rate": 4.728773584905661e-05,
      "loss": 0.5871,
      "num_input_tokens_seen": 2365264,
      "step": 3610
    },
    {
      "epoch": 1.8946540880503144,
      "grad_norm": 0.6589673161506653,
      "learning_rate": 4.735324947589099e-05,
      "loss": 0.7387,
      "num_input_tokens_seen": 2368080,
      "step": 3615
    },
    {
      "epoch": 1.8972746331236898,
      "grad_norm": 0.4364610016345978,
      "learning_rate": 4.7418763102725364e-05,
      "loss": 0.5293,
      "num_input_tokens_seen": 2371696,
      "step": 3620
    },
    {
      "epoch": 1.899895178197065,
      "grad_norm": 0.4690805971622467,
      "learning_rate": 4.7484276729559753e-05,
      "loss": 0.497,
      "num_input_tokens_seen": 2374160,
      "step": 3625
    },
    {
      "epoch": 1.9025157232704402,
      "grad_norm": 0.45772498846054077,
      "learning_rate": 4.754979035639413e-05,
      "loss": 0.4517,
      "num_input_tokens_seen": 2376656,
      "step": 3630
    },
    {
      "epoch": 1.9051362683438156,
      "grad_norm": 0.5049992799758911,
      "learning_rate": 4.761530398322851e-05,
      "loss": 0.4865,
      "num_input_tokens_seen": 2379856,
      "step": 3635
    },
    {
      "epoch": 1.9077568134171907,
      "grad_norm": 0.5556554198265076,
      "learning_rate": 4.7680817610062896e-05,
      "loss": 0.6122,
      "num_input_tokens_seen": 2382736,
      "step": 3640
    },
    {
      "epoch": 1.9103773584905661,
      "grad_norm": 0.6485047340393066,
      "learning_rate": 4.774633123689728e-05,
      "loss": 0.5033,
      "num_input_tokens_seen": 2386672,
      "step": 3645
    },
    {
      "epoch": 1.9129979035639413,
      "grad_norm": 0.3413850665092468,
      "learning_rate": 4.7811844863731655e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 2389296,
      "step": 3650
    },
    {
      "epoch": 1.9156184486373165,
      "grad_norm": 0.4397573173046112,
      "learning_rate": 4.787735849056604e-05,
      "loss": 0.6011,
      "num_input_tokens_seen": 2392080,
      "step": 3655
    },
    {
      "epoch": 1.9182389937106918,
      "grad_norm": 0.31642061471939087,
      "learning_rate": 4.794287211740042e-05,
      "loss": 0.4112,
      "num_input_tokens_seen": 2395952,
      "step": 3660
    },
    {
      "epoch": 1.9208595387840672,
      "grad_norm": 0.26771920919418335,
      "learning_rate": 4.8008385744234804e-05,
      "loss": 0.4443,
      "num_input_tokens_seen": 2398512,
      "step": 3665
    },
    {
      "epoch": 1.9234800838574424,
      "grad_norm": 0.43027207255363464,
      "learning_rate": 4.807389937106919e-05,
      "loss": 0.5893,
      "num_input_tokens_seen": 2401904,
      "step": 3670
    },
    {
      "epoch": 1.9261006289308176,
      "grad_norm": 0.3365291953086853,
      "learning_rate": 4.8139412997903564e-05,
      "loss": 0.5373,
      "num_input_tokens_seen": 2405584,
      "step": 3675
    },
    {
      "epoch": 1.9287211740041927,
      "grad_norm": 0.28879258036613464,
      "learning_rate": 4.8204926624737954e-05,
      "loss": 0.6088,
      "num_input_tokens_seen": 2409712,
      "step": 3680
    },
    {
      "epoch": 1.9313417190775681,
      "grad_norm": 0.3653244376182556,
      "learning_rate": 4.827044025157233e-05,
      "loss": 0.6513,
      "num_input_tokens_seen": 2412240,
      "step": 3685
    },
    {
      "epoch": 1.9339622641509435,
      "grad_norm": 0.4111684262752533,
      "learning_rate": 4.8335953878406706e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 2415056,
      "step": 3690
    },
    {
      "epoch": 1.9365828092243187,
      "grad_norm": 0.7658645510673523,
      "learning_rate": 4.8401467505241096e-05,
      "loss": 0.4813,
      "num_input_tokens_seen": 2419184,
      "step": 3695
    },
    {
      "epoch": 1.9392033542976939,
      "grad_norm": 0.7261778116226196,
      "learning_rate": 4.846698113207547e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 2422512,
      "step": 3700
    },
    {
      "epoch": 1.941823899371069,
      "grad_norm": 0.5599669218063354,
      "learning_rate": 4.8532494758909855e-05,
      "loss": 0.4906,
      "num_input_tokens_seen": 2425936,
      "step": 3705
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 0.8055518865585327,
      "learning_rate": 4.859800838574424e-05,
      "loss": 0.6244,
      "num_input_tokens_seen": 2428944,
      "step": 3710
    },
    {
      "epoch": 1.9470649895178198,
      "grad_norm": 0.47346922755241394,
      "learning_rate": 4.8663522012578615e-05,
      "loss": 0.4149,
      "num_input_tokens_seen": 2432080,
      "step": 3715
    },
    {
      "epoch": 1.949685534591195,
      "grad_norm": 0.4754215478897095,
      "learning_rate": 4.8729035639413e-05,
      "loss": 0.4741,
      "num_input_tokens_seen": 2434864,
      "step": 3720
    },
    {
      "epoch": 1.9523060796645701,
      "grad_norm": 0.49095267057418823,
      "learning_rate": 4.879454926624738e-05,
      "loss": 0.4881,
      "num_input_tokens_seen": 2437104,
      "step": 3725
    },
    {
      "epoch": 1.9549266247379455,
      "grad_norm": 0.5337731242179871,
      "learning_rate": 4.8860062893081764e-05,
      "loss": 0.6788,
      "num_input_tokens_seen": 2439792,
      "step": 3730
    },
    {
      "epoch": 1.9575471698113207,
      "grad_norm": 0.33999037742614746,
      "learning_rate": 4.892557651991615e-05,
      "loss": 0.6603,
      "num_input_tokens_seen": 2442320,
      "step": 3735
    },
    {
      "epoch": 1.960167714884696,
      "grad_norm": 0.4348808228969574,
      "learning_rate": 4.899109014675053e-05,
      "loss": 0.4452,
      "num_input_tokens_seen": 2445136,
      "step": 3740
    },
    {
      "epoch": 1.9627882599580713,
      "grad_norm": 0.35005074739456177,
      "learning_rate": 4.9056603773584906e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 2448080,
      "step": 3745
    },
    {
      "epoch": 1.9654088050314464,
      "grad_norm": 0.3938979506492615,
      "learning_rate": 4.912211740041929e-05,
      "loss": 0.5436,
      "num_input_tokens_seen": 2451376,
      "step": 3750
    },
    {
      "epoch": 1.9680293501048218,
      "grad_norm": 0.39738863706588745,
      "learning_rate": 4.918763102725367e-05,
      "loss": 0.6124,
      "num_input_tokens_seen": 2454672,
      "step": 3755
    },
    {
      "epoch": 1.9706498951781972,
      "grad_norm": 0.4552461504936218,
      "learning_rate": 4.925314465408805e-05,
      "loss": 0.4727,
      "num_input_tokens_seen": 2457712,
      "step": 3760
    },
    {
      "epoch": 1.9732704402515724,
      "grad_norm": 0.4709089994430542,
      "learning_rate": 4.931865828092244e-05,
      "loss": 0.4997,
      "num_input_tokens_seen": 2461936,
      "step": 3765
    },
    {
      "epoch": 1.9758909853249476,
      "grad_norm": 0.4204275906085968,
      "learning_rate": 4.9384171907756815e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 2465296,
      "step": 3770
    },
    {
      "epoch": 1.9785115303983227,
      "grad_norm": 0.5529329180717468,
      "learning_rate": 4.944968553459119e-05,
      "loss": 0.5119,
      "num_input_tokens_seen": 2469872,
      "step": 3775
    },
    {
      "epoch": 1.9811320754716981,
      "grad_norm": 0.5365079045295715,
      "learning_rate": 4.951519916142558e-05,
      "loss": 0.3987,
      "num_input_tokens_seen": 2473840,
      "step": 3780
    },
    {
      "epoch": 1.9837526205450735,
      "grad_norm": 0.6028404831886292,
      "learning_rate": 4.958071278825996e-05,
      "loss": 0.5356,
      "num_input_tokens_seen": 2476944,
      "step": 3785
    },
    {
      "epoch": 1.9863731656184487,
      "grad_norm": 0.4190744161605835,
      "learning_rate": 4.964622641509434e-05,
      "loss": 0.5089,
      "num_input_tokens_seen": 2479792,
      "step": 3790
    },
    {
      "epoch": 1.9889937106918238,
      "grad_norm": 0.5267384052276611,
      "learning_rate": 4.9711740041928724e-05,
      "loss": 0.6229,
      "num_input_tokens_seen": 2484464,
      "step": 3795
    },
    {
      "epoch": 1.991614255765199,
      "grad_norm": 0.6239827275276184,
      "learning_rate": 4.977725366876311e-05,
      "loss": 0.4753,
      "num_input_tokens_seen": 2487312,
      "step": 3800
    },
    {
      "epoch": 1.9942348008385744,
      "grad_norm": 0.4867786765098572,
      "learning_rate": 4.984276729559749e-05,
      "loss": 0.4979,
      "num_input_tokens_seen": 2490288,
      "step": 3805
    },
    {
      "epoch": 1.9968553459119498,
      "grad_norm": 0.5058794617652893,
      "learning_rate": 4.9908280922431866e-05,
      "loss": 0.5382,
      "num_input_tokens_seen": 2494128,
      "step": 3810
    },
    {
      "epoch": 1.999475890985325,
      "grad_norm": 0.3881987929344177,
      "learning_rate": 4.997379454926625e-05,
      "loss": 0.4842,
      "num_input_tokens_seen": 2496848,
      "step": 3815
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.5254608392715454,
      "eval_runtime": 9.7188,
      "eval_samples_per_second": 87.254,
      "eval_steps_per_second": 21.813,
      "num_input_tokens_seen": 2497016,
      "step": 3816
    },
    {
      "epoch": 2.0020964360587,
      "grad_norm": 0.47318390011787415,
      "learning_rate": 4.999999905865133e-05,
      "loss": 0.4114,
      "num_input_tokens_seen": 2499256,
      "step": 3820
    },
    {
      "epoch": 2.0047169811320753,
      "grad_norm": 0.5625536441802979,
      "learning_rate": 4.999999330596526e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 2502200,
      "step": 3825
    },
    {
      "epoch": 2.007337526205451,
      "grad_norm": 0.7050062417984009,
      "learning_rate": 4.9999982323565806e-05,
      "loss": 0.6759,
      "num_input_tokens_seen": 2505080,
      "step": 3830
    },
    {
      "epoch": 2.009958071278826,
      "grad_norm": 0.5848516821861267,
      "learning_rate": 4.999996611145527e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 2508216,
      "step": 3835
    },
    {
      "epoch": 2.0125786163522013,
      "grad_norm": 0.35800716280937195,
      "learning_rate": 4.9999944669637024e-05,
      "loss": 0.3816,
      "num_input_tokens_seen": 2511128,
      "step": 3840
    },
    {
      "epoch": 2.0151991614255764,
      "grad_norm": 0.4565981328487396,
      "learning_rate": 4.999991799811557e-05,
      "loss": 0.5221,
      "num_input_tokens_seen": 2514776,
      "step": 3845
    },
    {
      "epoch": 2.0178197064989516,
      "grad_norm": 0.7782171368598938,
      "learning_rate": 4.99998860968965e-05,
      "loss": 0.6653,
      "num_input_tokens_seen": 2517528,
      "step": 3850
    },
    {
      "epoch": 2.020440251572327,
      "grad_norm": 0.31216204166412354,
      "learning_rate": 4.999984896598646e-05,
      "loss": 0.5594,
      "num_input_tokens_seen": 2523032,
      "step": 3855
    },
    {
      "epoch": 2.0230607966457024,
      "grad_norm": 0.2930601239204407,
      "learning_rate": 4.999980660539323e-05,
      "loss": 0.4961,
      "num_input_tokens_seen": 2526200,
      "step": 3860
    },
    {
      "epoch": 2.0256813417190775,
      "grad_norm": 0.5397394299507141,
      "learning_rate": 4.999975901512567e-05,
      "loss": 0.6029,
      "num_input_tokens_seen": 2532760,
      "step": 3865
    },
    {
      "epoch": 2.0283018867924527,
      "grad_norm": 0.3225628435611725,
      "learning_rate": 4.999970619519374e-05,
      "loss": 0.4359,
      "num_input_tokens_seen": 2535640,
      "step": 3870
    },
    {
      "epoch": 2.030922431865828,
      "grad_norm": 0.30523595213890076,
      "learning_rate": 4.999964814560849e-05,
      "loss": 0.4878,
      "num_input_tokens_seen": 2539288,
      "step": 3875
    },
    {
      "epoch": 2.0335429769392035,
      "grad_norm": 0.8211117386817932,
      "learning_rate": 4.999958486638205e-05,
      "loss": 0.5952,
      "num_input_tokens_seen": 2541880,
      "step": 3880
    },
    {
      "epoch": 2.0361635220125787,
      "grad_norm": 0.39820751547813416,
      "learning_rate": 4.999951635752767e-05,
      "loss": 0.5028,
      "num_input_tokens_seen": 2545080,
      "step": 3885
    },
    {
      "epoch": 2.038784067085954,
      "grad_norm": 0.3752994239330292,
      "learning_rate": 4.999944261905968e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 2548376,
      "step": 3890
    },
    {
      "epoch": 2.041404612159329,
      "grad_norm": 0.4864736795425415,
      "learning_rate": 4.99993636509935e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 2550616,
      "step": 3895
    },
    {
      "epoch": 2.0440251572327046,
      "grad_norm": 0.40318572521209717,
      "learning_rate": 4.999927945334565e-05,
      "loss": 0.6078,
      "num_input_tokens_seen": 2555000,
      "step": 3900
    },
    {
      "epoch": 2.04664570230608,
      "grad_norm": 0.6470389366149902,
      "learning_rate": 4.9999190026133745e-05,
      "loss": 0.5222,
      "num_input_tokens_seen": 2558328,
      "step": 3905
    },
    {
      "epoch": 2.049266247379455,
      "grad_norm": 0.3725380003452301,
      "learning_rate": 4.9999095369376506e-05,
      "loss": 0.5975,
      "num_input_tokens_seen": 2561592,
      "step": 3910
    },
    {
      "epoch": 2.05188679245283,
      "grad_norm": 0.3867211639881134,
      "learning_rate": 4.99989954830937e-05,
      "loss": 0.3726,
      "num_input_tokens_seen": 2564472,
      "step": 3915
    },
    {
      "epoch": 2.0545073375262053,
      "grad_norm": 0.7058817148208618,
      "learning_rate": 4.9998890367306255e-05,
      "loss": 0.4852,
      "num_input_tokens_seen": 2567352,
      "step": 3920
    },
    {
      "epoch": 2.057127882599581,
      "grad_norm": 0.3226329982280731,
      "learning_rate": 4.999878002203614e-05,
      "loss": 0.3811,
      "num_input_tokens_seen": 2571448,
      "step": 3925
    },
    {
      "epoch": 2.059748427672956,
      "grad_norm": 0.5803183913230896,
      "learning_rate": 4.999866444730645e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 2574264,
      "step": 3930
    },
    {
      "epoch": 2.0623689727463312,
      "grad_norm": 0.6136952042579651,
      "learning_rate": 4.999854364314136e-05,
      "loss": 0.6265,
      "num_input_tokens_seen": 2577592,
      "step": 3935
    },
    {
      "epoch": 2.0649895178197064,
      "grad_norm": 0.6093463897705078,
      "learning_rate": 4.999841760956614e-05,
      "loss": 0.5258,
      "num_input_tokens_seen": 2581624,
      "step": 3940
    },
    {
      "epoch": 2.0676100628930816,
      "grad_norm": 0.6510100364685059,
      "learning_rate": 4.999828634660715e-05,
      "loss": 0.5254,
      "num_input_tokens_seen": 2584888,
      "step": 3945
    },
    {
      "epoch": 2.070230607966457,
      "grad_norm": 0.7580634951591492,
      "learning_rate": 4.9998149854291856e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 2587672,
      "step": 3950
    },
    {
      "epoch": 2.0728511530398324,
      "grad_norm": 0.4612693786621094,
      "learning_rate": 4.99980081326488e-05,
      "loss": 0.4933,
      "num_input_tokens_seen": 2590136,
      "step": 3955
    },
    {
      "epoch": 2.0754716981132075,
      "grad_norm": 0.5291822552680969,
      "learning_rate": 4.999786118170763e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 2593976,
      "step": 3960
    },
    {
      "epoch": 2.0780922431865827,
      "grad_norm": 0.5724464058876038,
      "learning_rate": 4.99977090014991e-05,
      "loss": 1.0137,
      "num_input_tokens_seen": 2597208,
      "step": 3965
    },
    {
      "epoch": 2.080712788259958,
      "grad_norm": 0.5428030490875244,
      "learning_rate": 4.9997551592055034e-05,
      "loss": 0.53,
      "num_input_tokens_seen": 2600536,
      "step": 3970
    },
    {
      "epoch": 2.0833333333333335,
      "grad_norm": 0.3931518793106079,
      "learning_rate": 4.999738895340836e-05,
      "loss": 0.4456,
      "num_input_tokens_seen": 2603608,
      "step": 3975
    },
    {
      "epoch": 2.0859538784067087,
      "grad_norm": 0.5865291357040405,
      "learning_rate": 4.999722108559311e-05,
      "loss": 0.5192,
      "num_input_tokens_seen": 2607288,
      "step": 3980
    },
    {
      "epoch": 2.088574423480084,
      "grad_norm": 0.41000795364379883,
      "learning_rate": 4.9997047988644385e-05,
      "loss": 0.6204,
      "num_input_tokens_seen": 2610392,
      "step": 3985
    },
    {
      "epoch": 2.091194968553459,
      "grad_norm": 0.3940235376358032,
      "learning_rate": 4.99968696625984e-05,
      "loss": 0.6499,
      "num_input_tokens_seen": 2613208,
      "step": 3990
    },
    {
      "epoch": 2.0938155136268346,
      "grad_norm": 0.3876796066761017,
      "learning_rate": 4.999668610749246e-05,
      "loss": 0.6378,
      "num_input_tokens_seen": 2616472,
      "step": 3995
    },
    {
      "epoch": 2.0964360587002098,
      "grad_norm": 0.3649873733520508,
      "learning_rate": 4.999649732336498e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 2620504,
      "step": 4000
    },
    {
      "epoch": 2.099056603773585,
      "grad_norm": 0.4579493999481201,
      "learning_rate": 4.999630331025542e-05,
      "loss": 0.5582,
      "num_input_tokens_seen": 2624632,
      "step": 4005
    },
    {
      "epoch": 2.10167714884696,
      "grad_norm": 0.6012643575668335,
      "learning_rate": 4.9996104068204385e-05,
      "loss": 0.5083,
      "num_input_tokens_seen": 2627608,
      "step": 4010
    },
    {
      "epoch": 2.1042976939203353,
      "grad_norm": 0.2771188020706177,
      "learning_rate": 4.999589959725355e-05,
      "loss": 0.453,
      "num_input_tokens_seen": 2630232,
      "step": 4015
    },
    {
      "epoch": 2.106918238993711,
      "grad_norm": 0.3307882845401764,
      "learning_rate": 4.999568989744568e-05,
      "loss": 0.4499,
      "num_input_tokens_seen": 2633624,
      "step": 4020
    },
    {
      "epoch": 2.109538784067086,
      "grad_norm": 1.2728267908096313,
      "learning_rate": 4.999547496882466e-05,
      "loss": 0.5526,
      "num_input_tokens_seen": 2636280,
      "step": 4025
    },
    {
      "epoch": 2.1121593291404612,
      "grad_norm": 0.4708094894886017,
      "learning_rate": 4.999525481143543e-05,
      "loss": 0.5472,
      "num_input_tokens_seen": 2640152,
      "step": 4030
    },
    {
      "epoch": 2.1147798742138364,
      "grad_norm": 0.28338688611984253,
      "learning_rate": 4.999502942532406e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 2643800,
      "step": 4035
    },
    {
      "epoch": 2.1174004192872116,
      "grad_norm": 0.418192058801651,
      "learning_rate": 4.999479881053769e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 2647608,
      "step": 4040
    },
    {
      "epoch": 2.120020964360587,
      "grad_norm": 0.48465704917907715,
      "learning_rate": 4.999456296712457e-05,
      "loss": 0.5994,
      "num_input_tokens_seen": 2650200,
      "step": 4045
    },
    {
      "epoch": 2.1226415094339623,
      "grad_norm": 0.3638647496700287,
      "learning_rate": 4.9994321895134025e-05,
      "loss": 0.6322,
      "num_input_tokens_seen": 2653784,
      "step": 4050
    },
    {
      "epoch": 2.1252620545073375,
      "grad_norm": 0.25319093465805054,
      "learning_rate": 4.999407559461649e-05,
      "loss": 0.5293,
      "num_input_tokens_seen": 2658104,
      "step": 4055
    },
    {
      "epoch": 2.1278825995807127,
      "grad_norm": 0.6113107204437256,
      "learning_rate": 4.999382406562349e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 2661080,
      "step": 4060
    },
    {
      "epoch": 2.130503144654088,
      "grad_norm": 0.3790110945701599,
      "learning_rate": 4.999356730820765e-05,
      "loss": 0.4306,
      "num_input_tokens_seen": 2664312,
      "step": 4065
    },
    {
      "epoch": 2.1331236897274635,
      "grad_norm": 1.0641392469406128,
      "learning_rate": 4.999330532242266e-05,
      "loss": 0.5916,
      "num_input_tokens_seen": 2667704,
      "step": 4070
    },
    {
      "epoch": 2.1357442348008386,
      "grad_norm": 0.40626028180122375,
      "learning_rate": 4.999303810832333e-05,
      "loss": 0.413,
      "num_input_tokens_seen": 2671160,
      "step": 4075
    },
    {
      "epoch": 2.138364779874214,
      "grad_norm": 0.29755884408950806,
      "learning_rate": 4.999276566596557e-05,
      "loss": 0.5482,
      "num_input_tokens_seen": 2675544,
      "step": 4080
    },
    {
      "epoch": 2.140985324947589,
      "grad_norm": 0.545146107673645,
      "learning_rate": 4.9992487995406364e-05,
      "loss": 0.4904,
      "num_input_tokens_seen": 2678264,
      "step": 4085
    },
    {
      "epoch": 2.1436058700209646,
      "grad_norm": 0.5114283561706543,
      "learning_rate": 4.9992205096703806e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 2684120,
      "step": 4090
    },
    {
      "epoch": 2.1462264150943398,
      "grad_norm": 0.9364824891090393,
      "learning_rate": 4.9991916969917054e-05,
      "loss": 0.6851,
      "num_input_tokens_seen": 2686872,
      "step": 4095
    },
    {
      "epoch": 2.148846960167715,
      "grad_norm": 0.3736487329006195,
      "learning_rate": 4.99916236151064e-05,
      "loss": 0.4662,
      "num_input_tokens_seen": 2690456,
      "step": 4100
    },
    {
      "epoch": 2.15146750524109,
      "grad_norm": 0.393059641122818,
      "learning_rate": 4.999132503233321e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 2692824,
      "step": 4105
    },
    {
      "epoch": 2.1540880503144653,
      "grad_norm": 0.526392936706543,
      "learning_rate": 4.999102122165994e-05,
      "loss": 0.6145,
      "num_input_tokens_seen": 2695704,
      "step": 4110
    },
    {
      "epoch": 2.156708595387841,
      "grad_norm": 0.3430602252483368,
      "learning_rate": 4.9990712183150135e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 2699480,
      "step": 4115
    },
    {
      "epoch": 2.159329140461216,
      "grad_norm": 0.45212775468826294,
      "learning_rate": 4.9990397916868456e-05,
      "loss": 0.5699,
      "num_input_tokens_seen": 2703960,
      "step": 4120
    },
    {
      "epoch": 2.161949685534591,
      "grad_norm": 0.41828805208206177,
      "learning_rate": 4.999007842288064e-05,
      "loss": 0.4718,
      "num_input_tokens_seen": 2706552,
      "step": 4125
    },
    {
      "epoch": 2.1645702306079664,
      "grad_norm": 0.40063774585723877,
      "learning_rate": 4.9989753701253514e-05,
      "loss": 0.5628,
      "num_input_tokens_seen": 2709368,
      "step": 4130
    },
    {
      "epoch": 2.1671907756813416,
      "grad_norm": 0.5619683265686035,
      "learning_rate": 4.998942375205502e-05,
      "loss": 0.562,
      "num_input_tokens_seen": 2712248,
      "step": 4135
    },
    {
      "epoch": 2.169811320754717,
      "grad_norm": 0.41904720664024353,
      "learning_rate": 4.998908857535416e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 2715448,
      "step": 4140
    },
    {
      "epoch": 2.1724318658280923,
      "grad_norm": 0.3313712179660797,
      "learning_rate": 4.9988748171221065e-05,
      "loss": 0.4006,
      "num_input_tokens_seen": 2719064,
      "step": 4145
    },
    {
      "epoch": 2.1750524109014675,
      "grad_norm": 0.32533302903175354,
      "learning_rate": 4.998840253972695e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 2722488,
      "step": 4150
    },
    {
      "epoch": 2.1776729559748427,
      "grad_norm": 0.5774205923080444,
      "learning_rate": 4.9988051680944094e-05,
      "loss": 0.4232,
      "num_input_tokens_seen": 2725528,
      "step": 4155
    },
    {
      "epoch": 2.180293501048218,
      "grad_norm": 0.3149208724498749,
      "learning_rate": 4.998769559494591e-05,
      "loss": 0.4428,
      "num_input_tokens_seen": 2728824,
      "step": 4160
    },
    {
      "epoch": 2.1829140461215935,
      "grad_norm": 0.34817227721214294,
      "learning_rate": 4.998733428180689e-05,
      "loss": 0.6245,
      "num_input_tokens_seen": 2732472,
      "step": 4165
    },
    {
      "epoch": 2.1855345911949686,
      "grad_norm": 0.2669568359851837,
      "learning_rate": 4.998696774160259e-05,
      "loss": 0.51,
      "num_input_tokens_seen": 2735544,
      "step": 4170
    },
    {
      "epoch": 2.188155136268344,
      "grad_norm": 0.31509846448898315,
      "learning_rate": 4.9986595974409714e-05,
      "loss": 0.5195,
      "num_input_tokens_seen": 2739032,
      "step": 4175
    },
    {
      "epoch": 2.190775681341719,
      "grad_norm": 0.5233045816421509,
      "learning_rate": 4.998621898030603e-05,
      "loss": 0.6478,
      "num_input_tokens_seen": 2742456,
      "step": 4180
    },
    {
      "epoch": 2.1933962264150946,
      "grad_norm": 0.5533230304718018,
      "learning_rate": 4.998583675937039e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 2744888,
      "step": 4185
    },
    {
      "epoch": 2.1960167714884697,
      "grad_norm": 0.5431790947914124,
      "learning_rate": 4.998544931168274e-05,
      "loss": 0.6077,
      "num_input_tokens_seen": 2747512,
      "step": 4190
    },
    {
      "epoch": 2.198637316561845,
      "grad_norm": 0.46978703141212463,
      "learning_rate": 4.9985056637324154e-05,
      "loss": 0.5285,
      "num_input_tokens_seen": 2750648,
      "step": 4195
    },
    {
      "epoch": 2.20125786163522,
      "grad_norm": 0.42163848876953125,
      "learning_rate": 4.998465873637677e-05,
      "loss": 0.402,
      "num_input_tokens_seen": 2753368,
      "step": 4200
    },
    {
      "epoch": 2.2038784067085953,
      "grad_norm": 0.42930692434310913,
      "learning_rate": 4.9984255608923806e-05,
      "loss": 0.4666,
      "num_input_tokens_seen": 2756696,
      "step": 4205
    },
    {
      "epoch": 2.2064989517819704,
      "grad_norm": 0.7173019647598267,
      "learning_rate": 4.998384725504961e-05,
      "loss": 0.536,
      "num_input_tokens_seen": 2759256,
      "step": 4210
    },
    {
      "epoch": 2.209119496855346,
      "grad_norm": 0.6400533318519592,
      "learning_rate": 4.9983433674839596e-05,
      "loss": 0.548,
      "num_input_tokens_seen": 2762296,
      "step": 4215
    },
    {
      "epoch": 2.211740041928721,
      "grad_norm": 0.7101878523826599,
      "learning_rate": 4.998301486838028e-05,
      "loss": 0.4835,
      "num_input_tokens_seen": 2765400,
      "step": 4220
    },
    {
      "epoch": 2.2143605870020964,
      "grad_norm": 1.4186060428619385,
      "learning_rate": 4.998259083575929e-05,
      "loss": 0.5037,
      "num_input_tokens_seen": 2768120,
      "step": 4225
    },
    {
      "epoch": 2.2169811320754715,
      "grad_norm": 0.4525040090084076,
      "learning_rate": 4.99821615770653e-05,
      "loss": 0.5389,
      "num_input_tokens_seen": 2771096,
      "step": 4230
    },
    {
      "epoch": 2.219601677148847,
      "grad_norm": 0.5576611161231995,
      "learning_rate": 4.998172709238813e-05,
      "loss": 0.582,
      "num_input_tokens_seen": 2774328,
      "step": 4235
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 0.36489689350128174,
      "learning_rate": 4.9981287381818654e-05,
      "loss": 0.5628,
      "num_input_tokens_seen": 2778008,
      "step": 4240
    },
    {
      "epoch": 2.2248427672955975,
      "grad_norm": 2.0153648853302,
      "learning_rate": 4.9980842445448856e-05,
      "loss": 0.5686,
      "num_input_tokens_seen": 2780216,
      "step": 4245
    },
    {
      "epoch": 2.2274633123689727,
      "grad_norm": 0.3843584358692169,
      "learning_rate": 4.9980392283371815e-05,
      "loss": 0.492,
      "num_input_tokens_seen": 2784088,
      "step": 4250
    },
    {
      "epoch": 2.230083857442348,
      "grad_norm": 0.32153916358947754,
      "learning_rate": 4.99799368956817e-05,
      "loss": 0.4217,
      "num_input_tokens_seen": 2786488,
      "step": 4255
    },
    {
      "epoch": 2.2327044025157234,
      "grad_norm": 0.5607828497886658,
      "learning_rate": 4.997947628247378e-05,
      "loss": 0.4375,
      "num_input_tokens_seen": 2788920,
      "step": 4260
    },
    {
      "epoch": 2.2353249475890986,
      "grad_norm": 0.5287513732910156,
      "learning_rate": 4.9979010443844395e-05,
      "loss": 0.6799,
      "num_input_tokens_seen": 2791640,
      "step": 4265
    },
    {
      "epoch": 2.237945492662474,
      "grad_norm": 0.42519983649253845,
      "learning_rate": 4.9978539379891e-05,
      "loss": 0.4198,
      "num_input_tokens_seen": 2794904,
      "step": 4270
    },
    {
      "epoch": 2.240566037735849,
      "grad_norm": 0.4007718861103058,
      "learning_rate": 4.997806309071214e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 2798840,
      "step": 4275
    },
    {
      "epoch": 2.243186582809224,
      "grad_norm": 0.42568618059158325,
      "learning_rate": 4.997758157640745e-05,
      "loss": 0.5564,
      "num_input_tokens_seen": 2802616,
      "step": 4280
    },
    {
      "epoch": 2.2458071278825997,
      "grad_norm": 0.5954672694206238,
      "learning_rate": 4.997709483707765e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 2805432,
      "step": 4285
    },
    {
      "epoch": 2.248427672955975,
      "grad_norm": 0.7823678255081177,
      "learning_rate": 4.997660287282456e-05,
      "loss": 0.5417,
      "num_input_tokens_seen": 2808696,
      "step": 4290
    },
    {
      "epoch": 2.25104821802935,
      "grad_norm": 0.34758031368255615,
      "learning_rate": 4.99761056837511e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 2811640,
      "step": 4295
    },
    {
      "epoch": 2.2536687631027252,
      "grad_norm": 0.4667873978614807,
      "learning_rate": 4.997560326996128e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 2815224,
      "step": 4300
    },
    {
      "epoch": 2.2562893081761004,
      "grad_norm": 0.4672618806362152,
      "learning_rate": 4.997509563156018e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 2817688,
      "step": 4305
    },
    {
      "epoch": 2.258909853249476,
      "grad_norm": 0.31105098128318787,
      "learning_rate": 4.997458276865401e-05,
      "loss": 0.4239,
      "num_input_tokens_seen": 2821016,
      "step": 4310
    },
    {
      "epoch": 2.261530398322851,
      "grad_norm": 0.4486381411552429,
      "learning_rate": 4.997406468135005e-05,
      "loss": 0.5947,
      "num_input_tokens_seen": 2823608,
      "step": 4315
    },
    {
      "epoch": 2.2641509433962264,
      "grad_norm": 0.4248979389667511,
      "learning_rate": 4.9973541369756686e-05,
      "loss": 0.4356,
      "num_input_tokens_seen": 2828248,
      "step": 4320
    },
    {
      "epoch": 2.2667714884696015,
      "grad_norm": 0.4668697416782379,
      "learning_rate": 4.9973012833983365e-05,
      "loss": 0.5262,
      "num_input_tokens_seen": 2832344,
      "step": 4325
    },
    {
      "epoch": 2.269392033542977,
      "grad_norm": 0.5176926851272583,
      "learning_rate": 4.9972479074140684e-05,
      "loss": 0.4548,
      "num_input_tokens_seen": 2836408,
      "step": 4330
    },
    {
      "epoch": 2.2720125786163523,
      "grad_norm": 0.3080736994743347,
      "learning_rate": 4.9971940090340276e-05,
      "loss": 0.5356,
      "num_input_tokens_seen": 2839320,
      "step": 4335
    },
    {
      "epoch": 2.2746331236897275,
      "grad_norm": 0.8074634671211243,
      "learning_rate": 4.997139588269489e-05,
      "loss": 0.5786,
      "num_input_tokens_seen": 2842456,
      "step": 4340
    },
    {
      "epoch": 2.2772536687631026,
      "grad_norm": 0.7256703972816467,
      "learning_rate": 4.997084645131838e-05,
      "loss": 0.6543,
      "num_input_tokens_seen": 2845432,
      "step": 4345
    },
    {
      "epoch": 2.279874213836478,
      "grad_norm": 0.3564213216304779,
      "learning_rate": 4.9970291796325666e-05,
      "loss": 0.5527,
      "num_input_tokens_seen": 2848536,
      "step": 4350
    },
    {
      "epoch": 2.2824947589098534,
      "grad_norm": 0.3769755959510803,
      "learning_rate": 4.996973191783279e-05,
      "loss": 0.5795,
      "num_input_tokens_seen": 2851544,
      "step": 4355
    },
    {
      "epoch": 2.2851153039832286,
      "grad_norm": 0.4179188907146454,
      "learning_rate": 4.996916681595687e-05,
      "loss": 0.4897,
      "num_input_tokens_seen": 2854392,
      "step": 4360
    },
    {
      "epoch": 2.2877358490566038,
      "grad_norm": 0.26464876532554626,
      "learning_rate": 4.996859649081611e-05,
      "loss": 0.4409,
      "num_input_tokens_seen": 2858424,
      "step": 4365
    },
    {
      "epoch": 2.290356394129979,
      "grad_norm": 0.3815103769302368,
      "learning_rate": 4.996802094252983e-05,
      "loss": 0.5263,
      "num_input_tokens_seen": 2861368,
      "step": 4370
    },
    {
      "epoch": 2.2929769392033545,
      "grad_norm": 0.7070125937461853,
      "learning_rate": 4.996744017121841e-05,
      "loss": 0.5435,
      "num_input_tokens_seen": 2863960,
      "step": 4375
    },
    {
      "epoch": 2.2955974842767297,
      "grad_norm": 0.8431401252746582,
      "learning_rate": 4.9966854177003355e-05,
      "loss": 0.5797,
      "num_input_tokens_seen": 2867352,
      "step": 4380
    },
    {
      "epoch": 2.298218029350105,
      "grad_norm": 0.6348665356636047,
      "learning_rate": 4.996626296000724e-05,
      "loss": 0.5537,
      "num_input_tokens_seen": 2871992,
      "step": 4385
    },
    {
      "epoch": 2.30083857442348,
      "grad_norm": 0.3998637795448303,
      "learning_rate": 4.996566652035374e-05,
      "loss": 0.4582,
      "num_input_tokens_seen": 2874680,
      "step": 4390
    },
    {
      "epoch": 2.3034591194968552,
      "grad_norm": 0.8725939989089966,
      "learning_rate": 4.996506485816763e-05,
      "loss": 0.3749,
      "num_input_tokens_seen": 2877688,
      "step": 4395
    },
    {
      "epoch": 2.3060796645702304,
      "grad_norm": 0.38620924949645996,
      "learning_rate": 4.996445797357477e-05,
      "loss": 0.5777,
      "num_input_tokens_seen": 2885528,
      "step": 4400
    },
    {
      "epoch": 2.308700209643606,
      "grad_norm": 0.372715562582016,
      "learning_rate": 4.996384586670211e-05,
      "loss": 0.6251,
      "num_input_tokens_seen": 2889464,
      "step": 4405
    },
    {
      "epoch": 2.311320754716981,
      "grad_norm": 0.5147207975387573,
      "learning_rate": 4.99632285376777e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 2893464,
      "step": 4410
    },
    {
      "epoch": 2.3139412997903563,
      "grad_norm": 0.5284803509712219,
      "learning_rate": 4.9962605986630674e-05,
      "loss": 0.5508,
      "num_input_tokens_seen": 2896792,
      "step": 4415
    },
    {
      "epoch": 2.3165618448637315,
      "grad_norm": 0.5985344052314758,
      "learning_rate": 4.996197821369126e-05,
      "loss": 0.5045,
      "num_input_tokens_seen": 2900664,
      "step": 4420
    },
    {
      "epoch": 2.319182389937107,
      "grad_norm": 0.5558983087539673,
      "learning_rate": 4.9961345218990796e-05,
      "loss": 0.5643,
      "num_input_tokens_seen": 2904152,
      "step": 4425
    },
    {
      "epoch": 2.3218029350104823,
      "grad_norm": 0.3913199007511139,
      "learning_rate": 4.996070700266168e-05,
      "loss": 0.5172,
      "num_input_tokens_seen": 2906776,
      "step": 4430
    },
    {
      "epoch": 2.3244234800838575,
      "grad_norm": 0.3737562596797943,
      "learning_rate": 4.9960063564837424e-05,
      "loss": 0.532,
      "num_input_tokens_seen": 2909720,
      "step": 4435
    },
    {
      "epoch": 2.3270440251572326,
      "grad_norm": 0.6617300510406494,
      "learning_rate": 4.995941490565262e-05,
      "loss": 0.3623,
      "num_input_tokens_seen": 2916536,
      "step": 4440
    },
    {
      "epoch": 2.329664570230608,
      "grad_norm": 0.5626647472381592,
      "learning_rate": 4.995876102524299e-05,
      "loss": 0.5639,
      "num_input_tokens_seen": 2919608,
      "step": 4445
    },
    {
      "epoch": 2.3322851153039834,
      "grad_norm": 0.35203856229782104,
      "learning_rate": 4.995810192374529e-05,
      "loss": 0.6426,
      "num_input_tokens_seen": 2923352,
      "step": 4450
    },
    {
      "epoch": 2.3349056603773586,
      "grad_norm": 0.49194467067718506,
      "learning_rate": 4.99574376012974e-05,
      "loss": 0.4984,
      "num_input_tokens_seen": 2926232,
      "step": 4455
    },
    {
      "epoch": 2.3375262054507338,
      "grad_norm": 0.5639827847480774,
      "learning_rate": 4.995676805803829e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 2929656,
      "step": 4460
    },
    {
      "epoch": 2.340146750524109,
      "grad_norm": 0.5175672769546509,
      "learning_rate": 4.995609329410804e-05,
      "loss": 0.5731,
      "num_input_tokens_seen": 2932344,
      "step": 4465
    },
    {
      "epoch": 2.342767295597484,
      "grad_norm": 0.2887899577617645,
      "learning_rate": 4.9955413309647776e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 2936152,
      "step": 4470
    },
    {
      "epoch": 2.3453878406708597,
      "grad_norm": 0.444763720035553,
      "learning_rate": 4.9954728104799756e-05,
      "loss": 0.4636,
      "num_input_tokens_seen": 2939224,
      "step": 4475
    },
    {
      "epoch": 2.348008385744235,
      "grad_norm": 0.2928982079029083,
      "learning_rate": 4.995403767970731e-05,
      "loss": 0.4176,
      "num_input_tokens_seen": 2941912,
      "step": 4480
    },
    {
      "epoch": 2.35062893081761,
      "grad_norm": 0.2991534471511841,
      "learning_rate": 4.995334203451488e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 2944568,
      "step": 4485
    },
    {
      "epoch": 2.353249475890985,
      "grad_norm": 0.37845832109451294,
      "learning_rate": 4.995264116936798e-05,
      "loss": 0.412,
      "num_input_tokens_seen": 2947480,
      "step": 4490
    },
    {
      "epoch": 2.3558700209643604,
      "grad_norm": 0.48611241579055786,
      "learning_rate": 4.995193508441322e-05,
      "loss": 0.5408,
      "num_input_tokens_seen": 2951192,
      "step": 4495
    },
    {
      "epoch": 2.358490566037736,
      "grad_norm": 0.7394600510597229,
      "learning_rate": 4.99512237797983e-05,
      "loss": 0.5227,
      "num_input_tokens_seen": 2954648,
      "step": 4500
    },
    {
      "epoch": 2.361111111111111,
      "grad_norm": 0.30275022983551025,
      "learning_rate": 4.995050725567203e-05,
      "loss": 0.4423,
      "num_input_tokens_seen": 2958552,
      "step": 4505
    },
    {
      "epoch": 2.3637316561844863,
      "grad_norm": 0.39439189434051514,
      "learning_rate": 4.994978551218429e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 2962040,
      "step": 4510
    },
    {
      "epoch": 2.3663522012578615,
      "grad_norm": 0.47840121388435364,
      "learning_rate": 4.9949058549486066e-05,
      "loss": 0.4822,
      "num_input_tokens_seen": 2964824,
      "step": 4515
    },
    {
      "epoch": 2.368972746331237,
      "grad_norm": 0.3358195126056671,
      "learning_rate": 4.994832636772943e-05,
      "loss": 0.4645,
      "num_input_tokens_seen": 2967448,
      "step": 4520
    },
    {
      "epoch": 2.3715932914046123,
      "grad_norm": 0.4750019907951355,
      "learning_rate": 4.994758896706754e-05,
      "loss": 0.5883,
      "num_input_tokens_seen": 2971000,
      "step": 4525
    },
    {
      "epoch": 2.3742138364779874,
      "grad_norm": 0.6944053173065186,
      "learning_rate": 4.994684634765465e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 2973592,
      "step": 4530
    },
    {
      "epoch": 2.3768343815513626,
      "grad_norm": 0.35167789459228516,
      "learning_rate": 4.9946098509646125e-05,
      "loss": 0.5775,
      "num_input_tokens_seen": 2977560,
      "step": 4535
    },
    {
      "epoch": 2.379454926624738,
      "grad_norm": 0.45731067657470703,
      "learning_rate": 4.994534545319838e-05,
      "loss": 0.4938,
      "num_input_tokens_seen": 2980280,
      "step": 4540
    },
    {
      "epoch": 2.3820754716981134,
      "grad_norm": 0.543713390827179,
      "learning_rate": 4.994458717846896e-05,
      "loss": 0.4845,
      "num_input_tokens_seen": 2983352,
      "step": 4545
    },
    {
      "epoch": 2.3846960167714886,
      "grad_norm": 0.5170190930366516,
      "learning_rate": 4.9943823685616485e-05,
      "loss": 0.5455,
      "num_input_tokens_seen": 2986648,
      "step": 4550
    },
    {
      "epoch": 2.3873165618448637,
      "grad_norm": 0.7595333456993103,
      "learning_rate": 4.994305497480067e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 2989880,
      "step": 4555
    },
    {
      "epoch": 2.389937106918239,
      "grad_norm": 0.45469754934310913,
      "learning_rate": 4.994228104618233e-05,
      "loss": 0.4671,
      "num_input_tokens_seen": 2992536,
      "step": 4560
    },
    {
      "epoch": 2.392557651991614,
      "grad_norm": 1.084980845451355,
      "learning_rate": 4.994150189992334e-05,
      "loss": 0.5222,
      "num_input_tokens_seen": 2996152,
      "step": 4565
    },
    {
      "epoch": 2.3951781970649897,
      "grad_norm": 1.0693812370300293,
      "learning_rate": 4.99407175361867e-05,
      "loss": 0.4695,
      "num_input_tokens_seen": 2999288,
      "step": 4570
    },
    {
      "epoch": 2.397798742138365,
      "grad_norm": 0.5579112768173218,
      "learning_rate": 4.99399279551365e-05,
      "loss": 0.3881,
      "num_input_tokens_seen": 3002520,
      "step": 4575
    },
    {
      "epoch": 2.40041928721174,
      "grad_norm": 0.4247887432575226,
      "learning_rate": 4.9939133156937886e-05,
      "loss": 0.5786,
      "num_input_tokens_seen": 3005752,
      "step": 4580
    },
    {
      "epoch": 2.403039832285115,
      "grad_norm": 0.5261085629463196,
      "learning_rate": 4.993833314175716e-05,
      "loss": 0.5178,
      "num_input_tokens_seen": 3009432,
      "step": 4585
    },
    {
      "epoch": 2.4056603773584904,
      "grad_norm": 0.36674395203590393,
      "learning_rate": 4.993752790976163e-05,
      "loss": 0.5239,
      "num_input_tokens_seen": 3014136,
      "step": 4590
    },
    {
      "epoch": 2.408280922431866,
      "grad_norm": 0.3977576494216919,
      "learning_rate": 4.993671746111977e-05,
      "loss": 0.5456,
      "num_input_tokens_seen": 3017144,
      "step": 4595
    },
    {
      "epoch": 2.410901467505241,
      "grad_norm": 5.9723219871521,
      "learning_rate": 4.993590179600111e-05,
      "loss": 0.6091,
      "num_input_tokens_seen": 3019448,
      "step": 4600
    },
    {
      "epoch": 2.4135220125786163,
      "grad_norm": 0.4582720398902893,
      "learning_rate": 4.9935080914576284e-05,
      "loss": 0.3986,
      "num_input_tokens_seen": 3022104,
      "step": 4605
    },
    {
      "epoch": 2.4161425576519915,
      "grad_norm": 0.4194348454475403,
      "learning_rate": 4.9934254817016995e-05,
      "loss": 0.5503,
      "num_input_tokens_seen": 3025688,
      "step": 4610
    },
    {
      "epoch": 2.418763102725367,
      "grad_norm": 0.3678266108036041,
      "learning_rate": 4.993342350349607e-05,
      "loss": 0.3715,
      "num_input_tokens_seen": 3029816,
      "step": 4615
    },
    {
      "epoch": 2.4213836477987423,
      "grad_norm": 0.3153185546398163,
      "learning_rate": 4.9932586974187404e-05,
      "loss": 0.3558,
      "num_input_tokens_seen": 3032824,
      "step": 4620
    },
    {
      "epoch": 2.4240041928721174,
      "grad_norm": 0.45628154277801514,
      "learning_rate": 4.9931745229265986e-05,
      "loss": 0.6242,
      "num_input_tokens_seen": 3036632,
      "step": 4625
    },
    {
      "epoch": 2.4266247379454926,
      "grad_norm": 0.6617905497550964,
      "learning_rate": 4.99308982689079e-05,
      "loss": 0.5399,
      "num_input_tokens_seen": 3039384,
      "step": 4630
    },
    {
      "epoch": 2.4292452830188678,
      "grad_norm": 0.444614052772522,
      "learning_rate": 4.993004609329033e-05,
      "loss": 0.506,
      "num_input_tokens_seen": 3042296,
      "step": 4635
    },
    {
      "epoch": 2.431865828092243,
      "grad_norm": 0.4636668860912323,
      "learning_rate": 4.992918870259153e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 3045304,
      "step": 4640
    },
    {
      "epoch": 2.4344863731656186,
      "grad_norm": 0.5095797777175903,
      "learning_rate": 4.992832609699086e-05,
      "loss": 0.4074,
      "num_input_tokens_seen": 3048824,
      "step": 4645
    },
    {
      "epoch": 2.4371069182389937,
      "grad_norm": 0.3453007638454437,
      "learning_rate": 4.992745827666877e-05,
      "loss": 0.5577,
      "num_input_tokens_seen": 3052120,
      "step": 4650
    },
    {
      "epoch": 2.439727463312369,
      "grad_norm": 0.29873624444007874,
      "learning_rate": 4.99265852418068e-05,
      "loss": 0.5086,
      "num_input_tokens_seen": 3055384,
      "step": 4655
    },
    {
      "epoch": 2.442348008385744,
      "grad_norm": 0.41098734736442566,
      "learning_rate": 4.992570699258757e-05,
      "loss": 0.402,
      "num_input_tokens_seen": 3058328,
      "step": 4660
    },
    {
      "epoch": 2.4449685534591197,
      "grad_norm": 0.3846357464790344,
      "learning_rate": 4.99248235291948e-05,
      "loss": 0.5533,
      "num_input_tokens_seen": 3063512,
      "step": 4665
    },
    {
      "epoch": 2.447589098532495,
      "grad_norm": 0.47352585196495056,
      "learning_rate": 4.9923934851813314e-05,
      "loss": 0.5314,
      "num_input_tokens_seen": 3068056,
      "step": 4670
    },
    {
      "epoch": 2.45020964360587,
      "grad_norm": 0.4428907334804535,
      "learning_rate": 4.9923040960628996e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 3071608,
      "step": 4675
    },
    {
      "epoch": 2.452830188679245,
      "grad_norm": 0.5927204489707947,
      "learning_rate": 4.9922141855828856e-05,
      "loss": 0.5166,
      "num_input_tokens_seen": 3074872,
      "step": 4680
    },
    {
      "epoch": 2.4554507337526204,
      "grad_norm": 0.4314022362232208,
      "learning_rate": 4.992123753760096e-05,
      "loss": 0.4474,
      "num_input_tokens_seen": 3077720,
      "step": 4685
    },
    {
      "epoch": 2.458071278825996,
      "grad_norm": 0.5098945498466492,
      "learning_rate": 4.9920328006134484e-05,
      "loss": 0.5955,
      "num_input_tokens_seen": 3080664,
      "step": 4690
    },
    {
      "epoch": 2.460691823899371,
      "grad_norm": 0.6710276007652283,
      "learning_rate": 4.991941326161971e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 3084408,
      "step": 4695
    },
    {
      "epoch": 2.4633123689727463,
      "grad_norm": 0.3490124046802521,
      "learning_rate": 4.991849330424796e-05,
      "loss": 0.5505,
      "num_input_tokens_seen": 3087640,
      "step": 4700
    },
    {
      "epoch": 2.4659329140461215,
      "grad_norm": 0.5001984238624573,
      "learning_rate": 4.9917568134211703e-05,
      "loss": 0.5836,
      "num_input_tokens_seen": 3090520,
      "step": 4705
    },
    {
      "epoch": 2.468553459119497,
      "grad_norm": 0.3492392301559448,
      "learning_rate": 4.991663775170447e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 3093624,
      "step": 4710
    },
    {
      "epoch": 2.4711740041928723,
      "grad_norm": 0.3880801796913147,
      "learning_rate": 4.991570215692088e-05,
      "loss": 0.445,
      "num_input_tokens_seen": 3096184,
      "step": 4715
    },
    {
      "epoch": 2.4737945492662474,
      "grad_norm": 0.48382654786109924,
      "learning_rate": 4.991476135005665e-05,
      "loss": 0.5906,
      "num_input_tokens_seen": 3099928,
      "step": 4720
    },
    {
      "epoch": 2.4764150943396226,
      "grad_norm": 0.7687484622001648,
      "learning_rate": 4.991381533130859e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 3102456,
      "step": 4725
    },
    {
      "epoch": 2.4790356394129978,
      "grad_norm": 1.8595105409622192,
      "learning_rate": 4.99128641008746e-05,
      "loss": 0.6168,
      "num_input_tokens_seen": 3104728,
      "step": 4730
    },
    {
      "epoch": 2.481656184486373,
      "grad_norm": 0.8053106069564819,
      "learning_rate": 4.9911907658953664e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 3107704,
      "step": 4735
    },
    {
      "epoch": 2.4842767295597485,
      "grad_norm": 0.38350850343704224,
      "learning_rate": 4.991094600574584e-05,
      "loss": 0.4387,
      "num_input_tokens_seen": 3110328,
      "step": 4740
    },
    {
      "epoch": 2.4868972746331237,
      "grad_norm": 0.4130067825317383,
      "learning_rate": 4.9909979141452324e-05,
      "loss": 0.5289,
      "num_input_tokens_seen": 3113464,
      "step": 4745
    },
    {
      "epoch": 2.489517819706499,
      "grad_norm": 0.2739948630332947,
      "learning_rate": 4.9909007066275346e-05,
      "loss": 0.425,
      "num_input_tokens_seen": 3117112,
      "step": 4750
    },
    {
      "epoch": 2.492138364779874,
      "grad_norm": 0.31161007285118103,
      "learning_rate": 4.990802978041828e-05,
      "loss": 0.4584,
      "num_input_tokens_seen": 3120280,
      "step": 4755
    },
    {
      "epoch": 2.4947589098532497,
      "grad_norm": 0.35810670256614685,
      "learning_rate": 4.990704728408555e-05,
      "loss": 0.4919,
      "num_input_tokens_seen": 3123928,
      "step": 4760
    },
    {
      "epoch": 2.497379454926625,
      "grad_norm": 0.3879152834415436,
      "learning_rate": 4.990605957748268e-05,
      "loss": 0.4142,
      "num_input_tokens_seen": 3126584,
      "step": 4765
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.6474513411521912,
      "learning_rate": 4.990506666081628e-05,
      "loss": 0.635,
      "num_input_tokens_seen": 3129368,
      "step": 4770
    },
    {
      "epoch": 2.502620545073375,
      "grad_norm": 0.9548970460891724,
      "learning_rate": 4.990406853429408e-05,
      "loss": 0.5704,
      "num_input_tokens_seen": 3132248,
      "step": 4775
    },
    {
      "epoch": 2.5052410901467503,
      "grad_norm": 0.34809744358062744,
      "learning_rate": 4.990306519812485e-05,
      "loss": 0.3761,
      "num_input_tokens_seen": 3135320,
      "step": 4780
    },
    {
      "epoch": 2.507861635220126,
      "grad_norm": 0.5991567373275757,
      "learning_rate": 4.99020566525185e-05,
      "loss": 0.706,
      "num_input_tokens_seen": 3139032,
      "step": 4785
    },
    {
      "epoch": 2.510482180293501,
      "grad_norm": 0.5188328623771667,
      "learning_rate": 4.9901042897685996e-05,
      "loss": 0.6627,
      "num_input_tokens_seen": 3141944,
      "step": 4790
    },
    {
      "epoch": 2.5131027253668763,
      "grad_norm": 0.9085689783096313,
      "learning_rate": 4.99000239338394e-05,
      "loss": 0.5343,
      "num_input_tokens_seen": 3145400,
      "step": 4795
    },
    {
      "epoch": 2.5157232704402515,
      "grad_norm": 0.3406851887702942,
      "learning_rate": 4.989899976119187e-05,
      "loss": 0.5935,
      "num_input_tokens_seen": 3149496,
      "step": 4800
    },
    {
      "epoch": 2.518343815513627,
      "grad_norm": 0.4125145971775055,
      "learning_rate": 4.989797037995766e-05,
      "loss": 0.6129,
      "num_input_tokens_seen": 3152344,
      "step": 4805
    },
    {
      "epoch": 2.5209643605870022,
      "grad_norm": 0.45011579990386963,
      "learning_rate": 4.989693579035209e-05,
      "loss": 0.418,
      "num_input_tokens_seen": 3155192,
      "step": 4810
    },
    {
      "epoch": 2.5235849056603774,
      "grad_norm": 0.4806981086730957,
      "learning_rate": 4.989589599259159e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 3157688,
      "step": 4815
    },
    {
      "epoch": 2.5262054507337526,
      "grad_norm": 0.2719520628452301,
      "learning_rate": 4.9894850986893685e-05,
      "loss": 0.4954,
      "num_input_tokens_seen": 3161592,
      "step": 4820
    },
    {
      "epoch": 2.5288259958071277,
      "grad_norm": 0.6537134051322937,
      "learning_rate": 4.9893800773476965e-05,
      "loss": 0.4836,
      "num_input_tokens_seen": 3164024,
      "step": 4825
    },
    {
      "epoch": 2.531446540880503,
      "grad_norm": 0.570635974407196,
      "learning_rate": 4.9892745352561134e-05,
      "loss": 0.4242,
      "num_input_tokens_seen": 3166424,
      "step": 4830
    },
    {
      "epoch": 2.5340670859538785,
      "grad_norm": 0.558092474937439,
      "learning_rate": 4.9891684724366955e-05,
      "loss": 0.6854,
      "num_input_tokens_seen": 3169240,
      "step": 4835
    },
    {
      "epoch": 2.5366876310272537,
      "grad_norm": 0.6869202256202698,
      "learning_rate": 4.989061888911631e-05,
      "loss": 0.5074,
      "num_input_tokens_seen": 3172088,
      "step": 4840
    },
    {
      "epoch": 2.539308176100629,
      "grad_norm": 0.5174260139465332,
      "learning_rate": 4.988954784703217e-05,
      "loss": 0.527,
      "num_input_tokens_seen": 3175480,
      "step": 4845
    },
    {
      "epoch": 2.541928721174004,
      "grad_norm": 0.4259769916534424,
      "learning_rate": 4.9888471598338574e-05,
      "loss": 0.5404,
      "num_input_tokens_seen": 3179480,
      "step": 4850
    },
    {
      "epoch": 2.5445492662473796,
      "grad_norm": 0.3050215542316437,
      "learning_rate": 4.988739014326066e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 3182520,
      "step": 4855
    },
    {
      "epoch": 2.547169811320755,
      "grad_norm": 0.5463882684707642,
      "learning_rate": 4.9886303482024656e-05,
      "loss": 0.5321,
      "num_input_tokens_seen": 3185560,
      "step": 4860
    },
    {
      "epoch": 2.54979035639413,
      "grad_norm": 0.3470593988895416,
      "learning_rate": 4.988521161485789e-05,
      "loss": 0.5041,
      "num_input_tokens_seen": 3189592,
      "step": 4865
    },
    {
      "epoch": 2.552410901467505,
      "grad_norm": 0.8342275023460388,
      "learning_rate": 4.9884114541988747e-05,
      "loss": 0.4254,
      "num_input_tokens_seen": 3192760,
      "step": 4870
    },
    {
      "epoch": 2.5550314465408803,
      "grad_norm": 0.810166597366333,
      "learning_rate": 4.988301226364674e-05,
      "loss": 0.4291,
      "num_input_tokens_seen": 3195992,
      "step": 4875
    },
    {
      "epoch": 2.5576519916142555,
      "grad_norm": 0.8341299295425415,
      "learning_rate": 4.988190478006245e-05,
      "loss": 0.5132,
      "num_input_tokens_seen": 3198200,
      "step": 4880
    },
    {
      "epoch": 2.560272536687631,
      "grad_norm": 0.3327268660068512,
      "learning_rate": 4.988079209146754e-05,
      "loss": 0.4774,
      "num_input_tokens_seen": 3201848,
      "step": 4885
    },
    {
      "epoch": 2.5628930817610063,
      "grad_norm": 1.2911386489868164,
      "learning_rate": 4.9879674198094794e-05,
      "loss": 0.666,
      "num_input_tokens_seen": 3204504,
      "step": 4890
    },
    {
      "epoch": 2.5655136268343814,
      "grad_norm": 0.678874671459198,
      "learning_rate": 4.987855110017804e-05,
      "loss": 0.6441,
      "num_input_tokens_seen": 3207320,
      "step": 4895
    },
    {
      "epoch": 2.568134171907757,
      "grad_norm": 0.4688921570777893,
      "learning_rate": 4.9877422797952224e-05,
      "loss": 0.4614,
      "num_input_tokens_seen": 3210584,
      "step": 4900
    },
    {
      "epoch": 2.5707547169811322,
      "grad_norm": 0.35241755843162537,
      "learning_rate": 4.987628929165338e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 3213496,
      "step": 4905
    },
    {
      "epoch": 2.5733752620545074,
      "grad_norm": 0.3082263171672821,
      "learning_rate": 4.9875150581518615e-05,
      "loss": 0.4829,
      "num_input_tokens_seen": 3217016,
      "step": 4910
    },
    {
      "epoch": 2.5759958071278826,
      "grad_norm": 0.3814629912376404,
      "learning_rate": 4.987400666778614e-05,
      "loss": 0.6282,
      "num_input_tokens_seen": 3219800,
      "step": 4915
    },
    {
      "epoch": 2.5786163522012577,
      "grad_norm": 0.6843031048774719,
      "learning_rate": 4.987285755069524e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 3222424,
      "step": 4920
    },
    {
      "epoch": 2.581236897274633,
      "grad_norm": 0.507969081401825,
      "learning_rate": 4.987170323048632e-05,
      "loss": 0.5041,
      "num_input_tokens_seen": 3225272,
      "step": 4925
    },
    {
      "epoch": 2.5838574423480085,
      "grad_norm": 0.2698114216327667,
      "learning_rate": 4.987054370740083e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 3228088,
      "step": 4930
    },
    {
      "epoch": 2.5864779874213837,
      "grad_norm": 0.3725367486476898,
      "learning_rate": 4.986937898168134e-05,
      "loss": 0.4892,
      "num_input_tokens_seen": 3231096,
      "step": 4935
    },
    {
      "epoch": 2.589098532494759,
      "grad_norm": 0.49054858088493347,
      "learning_rate": 4.986820905357148e-05,
      "loss": 0.5928,
      "num_input_tokens_seen": 3234552,
      "step": 4940
    },
    {
      "epoch": 2.591719077568134,
      "grad_norm": 0.6970625519752502,
      "learning_rate": 4.9867033923315995e-05,
      "loss": 0.4005,
      "num_input_tokens_seen": 3238072,
      "step": 4945
    },
    {
      "epoch": 2.5943396226415096,
      "grad_norm": 0.7218769788742065,
      "learning_rate": 4.9865853591160724e-05,
      "loss": 0.5632,
      "num_input_tokens_seen": 3241624,
      "step": 4950
    },
    {
      "epoch": 2.596960167714885,
      "grad_norm": 0.39386147260665894,
      "learning_rate": 4.9864668057352556e-05,
      "loss": 0.7843,
      "num_input_tokens_seen": 3245080,
      "step": 4955
    },
    {
      "epoch": 2.59958071278826,
      "grad_norm": 0.4381054639816284,
      "learning_rate": 4.9863477322139517e-05,
      "loss": 0.5372,
      "num_input_tokens_seen": 3247800,
      "step": 4960
    },
    {
      "epoch": 2.602201257861635,
      "grad_norm": 0.31119292974472046,
      "learning_rate": 4.986228138577067e-05,
      "loss": 0.382,
      "num_input_tokens_seen": 3251000,
      "step": 4965
    },
    {
      "epoch": 2.6048218029350103,
      "grad_norm": 0.42748117446899414,
      "learning_rate": 4.986108024849619e-05,
      "loss": 0.4136,
      "num_input_tokens_seen": 3254872,
      "step": 4970
    },
    {
      "epoch": 2.6074423480083855,
      "grad_norm": 0.568141520023346,
      "learning_rate": 4.985987391056737e-05,
      "loss": 0.527,
      "num_input_tokens_seen": 3258520,
      "step": 4975
    },
    {
      "epoch": 2.610062893081761,
      "grad_norm": 0.4234444200992584,
      "learning_rate": 4.9858662372236544e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 3261944,
      "step": 4980
    },
    {
      "epoch": 2.6126834381551363,
      "grad_norm": 0.5293048620223999,
      "learning_rate": 4.985744563375714e-05,
      "loss": 0.5889,
      "num_input_tokens_seen": 3265432,
      "step": 4985
    },
    {
      "epoch": 2.6153039832285114,
      "grad_norm": 0.6991572380065918,
      "learning_rate": 4.985622369538371e-05,
      "loss": 0.5364,
      "num_input_tokens_seen": 3268632,
      "step": 4990
    },
    {
      "epoch": 2.617924528301887,
      "grad_norm": 0.8364831209182739,
      "learning_rate": 4.985499655737185e-05,
      "loss": 0.4702,
      "num_input_tokens_seen": 3271416,
      "step": 4995
    },
    {
      "epoch": 2.620545073375262,
      "grad_norm": 0.5066753029823303,
      "learning_rate": 4.985376421997827e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 3274424,
      "step": 5000
    },
    {
      "epoch": 2.6231656184486374,
      "grad_norm": 0.44661593437194824,
      "learning_rate": 4.985252668346076e-05,
      "loss": 0.4908,
      "num_input_tokens_seen": 3277688,
      "step": 5005
    },
    {
      "epoch": 2.6257861635220126,
      "grad_norm": 0.711195707321167,
      "learning_rate": 4.9851283948078206e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 3280440,
      "step": 5010
    },
    {
      "epoch": 2.6284067085953877,
      "grad_norm": 0.5105148553848267,
      "learning_rate": 4.985003601409057e-05,
      "loss": 0.671,
      "num_input_tokens_seen": 3283704,
      "step": 5015
    },
    {
      "epoch": 2.631027253668763,
      "grad_norm": 0.44942334294319153,
      "learning_rate": 4.9848782881758906e-05,
      "loss": 0.4582,
      "num_input_tokens_seen": 3286808,
      "step": 5020
    },
    {
      "epoch": 2.6336477987421385,
      "grad_norm": 0.3227616548538208,
      "learning_rate": 4.9847524551345345e-05,
      "loss": 0.5964,
      "num_input_tokens_seen": 3289592,
      "step": 5025
    },
    {
      "epoch": 2.6362683438155137,
      "grad_norm": 0.3208800256252289,
      "learning_rate": 4.984626102311313e-05,
      "loss": 0.5023,
      "num_input_tokens_seen": 3292696,
      "step": 5030
    },
    {
      "epoch": 2.638888888888889,
      "grad_norm": 0.5121229290962219,
      "learning_rate": 4.984499229732656e-05,
      "loss": 0.3514,
      "num_input_tokens_seen": 3296632,
      "step": 5035
    },
    {
      "epoch": 2.641509433962264,
      "grad_norm": 0.5023754239082336,
      "learning_rate": 4.984371837425106e-05,
      "loss": 0.4253,
      "num_input_tokens_seen": 3300536,
      "step": 5040
    },
    {
      "epoch": 2.6441299790356396,
      "grad_norm": 0.46870696544647217,
      "learning_rate": 4.9842439254153106e-05,
      "loss": 0.4863,
      "num_input_tokens_seen": 3304728,
      "step": 5045
    },
    {
      "epoch": 2.646750524109015,
      "grad_norm": 0.5894140005111694,
      "learning_rate": 4.984115493730027e-05,
      "loss": 0.5534,
      "num_input_tokens_seen": 3307928,
      "step": 5050
    },
    {
      "epoch": 2.64937106918239,
      "grad_norm": 0.6688151955604553,
      "learning_rate": 4.983986542396123e-05,
      "loss": 0.5784,
      "num_input_tokens_seen": 3310552,
      "step": 5055
    },
    {
      "epoch": 2.651991614255765,
      "grad_norm": 0.42543303966522217,
      "learning_rate": 4.983857071440573e-05,
      "loss": 0.6214,
      "num_input_tokens_seen": 3314904,
      "step": 5060
    },
    {
      "epoch": 2.6546121593291403,
      "grad_norm": 0.4296688735485077,
      "learning_rate": 4.9837270808904616e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 3320184,
      "step": 5065
    },
    {
      "epoch": 2.6572327044025155,
      "grad_norm": 0.6404249668121338,
      "learning_rate": 4.9835965707729807e-05,
      "loss": 0.4443,
      "num_input_tokens_seen": 3322904,
      "step": 5070
    },
    {
      "epoch": 2.659853249475891,
      "grad_norm": 0.3348119556903839,
      "learning_rate": 4.983465541115431e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 3325784,
      "step": 5075
    },
    {
      "epoch": 2.6624737945492662,
      "grad_norm": 0.37885212898254395,
      "learning_rate": 4.983333991945223e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 3328760,
      "step": 5080
    },
    {
      "epoch": 2.6650943396226414,
      "grad_norm": 0.6513246893882751,
      "learning_rate": 4.983201923289875e-05,
      "loss": 0.5344,
      "num_input_tokens_seen": 3331576,
      "step": 5085
    },
    {
      "epoch": 2.667714884696017,
      "grad_norm": 0.883923351764679,
      "learning_rate": 4.983069335177014e-05,
      "loss": 0.5678,
      "num_input_tokens_seen": 3333880,
      "step": 5090
    },
    {
      "epoch": 2.670335429769392,
      "grad_norm": 0.6822410225868225,
      "learning_rate": 4.982936227634378e-05,
      "loss": 0.5027,
      "num_input_tokens_seen": 3337880,
      "step": 5095
    },
    {
      "epoch": 2.6729559748427674,
      "grad_norm": 0.29517677426338196,
      "learning_rate": 4.982802600689809e-05,
      "loss": 0.4919,
      "num_input_tokens_seen": 3340920,
      "step": 5100
    },
    {
      "epoch": 2.6755765199161425,
      "grad_norm": 0.6086233854293823,
      "learning_rate": 4.982668454371261e-05,
      "loss": 0.4712,
      "num_input_tokens_seen": 3344312,
      "step": 5105
    },
    {
      "epoch": 2.6781970649895177,
      "grad_norm": 0.5453270077705383,
      "learning_rate": 4.9825337887067966e-05,
      "loss": 0.437,
      "num_input_tokens_seen": 3347672,
      "step": 5110
    },
    {
      "epoch": 2.680817610062893,
      "grad_norm": 0.3542221486568451,
      "learning_rate": 4.982398603724586e-05,
      "loss": 0.5109,
      "num_input_tokens_seen": 3351032,
      "step": 5115
    },
    {
      "epoch": 2.6834381551362685,
      "grad_norm": 0.7338021993637085,
      "learning_rate": 4.982262899452909e-05,
      "loss": 0.6628,
      "num_input_tokens_seen": 3353752,
      "step": 5120
    },
    {
      "epoch": 2.6860587002096437,
      "grad_norm": 0.5757216811180115,
      "learning_rate": 4.982126675920151e-05,
      "loss": 0.4508,
      "num_input_tokens_seen": 3356600,
      "step": 5125
    },
    {
      "epoch": 2.688679245283019,
      "grad_norm": 0.9681112170219421,
      "learning_rate": 4.98198993315481e-05,
      "loss": 0.552,
      "num_input_tokens_seen": 3360280,
      "step": 5130
    },
    {
      "epoch": 2.691299790356394,
      "grad_norm": 0.3666017949581146,
      "learning_rate": 4.981852671185492e-05,
      "loss": 0.6144,
      "num_input_tokens_seen": 3363768,
      "step": 5135
    },
    {
      "epoch": 2.6939203354297696,
      "grad_norm": 0.411094069480896,
      "learning_rate": 4.9817148900409093e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 3367480,
      "step": 5140
    },
    {
      "epoch": 2.6965408805031448,
      "grad_norm": 0.7031425833702087,
      "learning_rate": 4.981576589749885e-05,
      "loss": 0.5042,
      "num_input_tokens_seen": 3370520,
      "step": 5145
    },
    {
      "epoch": 2.69916142557652,
      "grad_norm": 0.7773240208625793,
      "learning_rate": 4.981437770341348e-05,
      "loss": 0.4525,
      "num_input_tokens_seen": 3372984,
      "step": 5150
    },
    {
      "epoch": 2.701781970649895,
      "grad_norm": 1.3006759881973267,
      "learning_rate": 4.9812984318443414e-05,
      "loss": 0.4876,
      "num_input_tokens_seen": 3375288,
      "step": 5155
    },
    {
      "epoch": 2.7044025157232703,
      "grad_norm": 0.5373823046684265,
      "learning_rate": 4.98115857428801e-05,
      "loss": 0.4127,
      "num_input_tokens_seen": 3378488,
      "step": 5160
    },
    {
      "epoch": 2.7070230607966455,
      "grad_norm": 0.3480692505836487,
      "learning_rate": 4.9810181977016104e-05,
      "loss": 0.4506,
      "num_input_tokens_seen": 3381368,
      "step": 5165
    },
    {
      "epoch": 2.709643605870021,
      "grad_norm": 1.017441987991333,
      "learning_rate": 4.9808773021145096e-05,
      "loss": 0.5013,
      "num_input_tokens_seen": 3385016,
      "step": 5170
    },
    {
      "epoch": 2.7122641509433962,
      "grad_norm": 0.3272978365421295,
      "learning_rate": 4.98073588755618e-05,
      "loss": 0.5442,
      "num_input_tokens_seen": 3387992,
      "step": 5175
    },
    {
      "epoch": 2.7148846960167714,
      "grad_norm": 0.4389372766017914,
      "learning_rate": 4.980593954056205e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 3390872,
      "step": 5180
    },
    {
      "epoch": 2.717505241090147,
      "grad_norm": 0.4383937120437622,
      "learning_rate": 4.9804515016442754e-05,
      "loss": 0.5837,
      "num_input_tokens_seen": 3393496,
      "step": 5185
    },
    {
      "epoch": 2.720125786163522,
      "grad_norm": 0.7698715329170227,
      "learning_rate": 4.9803085303501896e-05,
      "loss": 0.7379,
      "num_input_tokens_seen": 3395992,
      "step": 5190
    },
    {
      "epoch": 2.7227463312368974,
      "grad_norm": 0.4678775370121002,
      "learning_rate": 4.9801650402038555e-05,
      "loss": 0.6021,
      "num_input_tokens_seen": 3399160,
      "step": 5195
    },
    {
      "epoch": 2.7253668763102725,
      "grad_norm": 0.5092703700065613,
      "learning_rate": 4.980021031235291e-05,
      "loss": 0.4969,
      "num_input_tokens_seen": 3402072,
      "step": 5200
    },
    {
      "epoch": 2.7279874213836477,
      "grad_norm": 0.6574476361274719,
      "learning_rate": 4.9798765034746206e-05,
      "loss": 0.5012,
      "num_input_tokens_seen": 3404344,
      "step": 5205
    },
    {
      "epoch": 2.730607966457023,
      "grad_norm": 0.3641524612903595,
      "learning_rate": 4.979731456952077e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 3407640,
      "step": 5210
    },
    {
      "epoch": 2.7332285115303985,
      "grad_norm": 0.702442467212677,
      "learning_rate": 4.9795858916980034e-05,
      "loss": 0.3914,
      "num_input_tokens_seen": 3410200,
      "step": 5215
    },
    {
      "epoch": 2.7358490566037736,
      "grad_norm": 0.46065956354141235,
      "learning_rate": 4.9794398077428495e-05,
      "loss": 0.4398,
      "num_input_tokens_seen": 3412856,
      "step": 5220
    },
    {
      "epoch": 2.738469601677149,
      "grad_norm": 0.25887200236320496,
      "learning_rate": 4.979293205117175e-05,
      "loss": 0.4666,
      "num_input_tokens_seen": 3415864,
      "step": 5225
    },
    {
      "epoch": 2.741090146750524,
      "grad_norm": 0.2879312336444855,
      "learning_rate": 4.979146083851647e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 3419224,
      "step": 5230
    },
    {
      "epoch": 2.7437106918238996,
      "grad_norm": 0.5751168727874756,
      "learning_rate": 4.978998443977042e-05,
      "loss": 0.4425,
      "num_input_tokens_seen": 3422520,
      "step": 5235
    },
    {
      "epoch": 2.7463312368972748,
      "grad_norm": 0.578065812587738,
      "learning_rate": 4.9788502855242436e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 3425976,
      "step": 5240
    },
    {
      "epoch": 2.74895178197065,
      "grad_norm": 0.4471054673194885,
      "learning_rate": 4.978701608524247e-05,
      "loss": 0.5018,
      "num_input_tokens_seen": 3429368,
      "step": 5245
    },
    {
      "epoch": 2.751572327044025,
      "grad_norm": 0.7011773586273193,
      "learning_rate": 4.978552413008152e-05,
      "loss": 0.5441,
      "num_input_tokens_seen": 3434680,
      "step": 5250
    },
    {
      "epoch": 2.7541928721174003,
      "grad_norm": 0.9564266204833984,
      "learning_rate": 4.9784026990071687e-05,
      "loss": 0.587,
      "num_input_tokens_seen": 3437752,
      "step": 5255
    },
    {
      "epoch": 2.7568134171907754,
      "grad_norm": 0.35832834243774414,
      "learning_rate": 4.978252466552615e-05,
      "loss": 0.6152,
      "num_input_tokens_seen": 3440760,
      "step": 5260
    },
    {
      "epoch": 2.759433962264151,
      "grad_norm": 0.558123767375946,
      "learning_rate": 4.9781017156759205e-05,
      "loss": 0.521,
      "num_input_tokens_seen": 3443736,
      "step": 5265
    },
    {
      "epoch": 2.762054507337526,
      "grad_norm": 0.9007126688957214,
      "learning_rate": 4.977950446408618e-05,
      "loss": 0.6285,
      "num_input_tokens_seen": 3446904,
      "step": 5270
    },
    {
      "epoch": 2.7646750524109014,
      "grad_norm": 0.4588097631931305,
      "learning_rate": 4.9777986587823523e-05,
      "loss": 0.4373,
      "num_input_tokens_seen": 3450104,
      "step": 5275
    },
    {
      "epoch": 2.767295597484277,
      "grad_norm": 0.3594784140586853,
      "learning_rate": 4.977646352828875e-05,
      "loss": 0.6619,
      "num_input_tokens_seen": 3453592,
      "step": 5280
    },
    {
      "epoch": 2.769916142557652,
      "grad_norm": 0.7781020402908325,
      "learning_rate": 4.9774935285800476e-05,
      "loss": 0.4971,
      "num_input_tokens_seen": 3456824,
      "step": 5285
    },
    {
      "epoch": 2.7725366876310273,
      "grad_norm": 0.7689031958580017,
      "learning_rate": 4.977340186067838e-05,
      "loss": 0.5253,
      "num_input_tokens_seen": 3460792,
      "step": 5290
    },
    {
      "epoch": 2.7751572327044025,
      "grad_norm": 0.3803282678127289,
      "learning_rate": 4.977186325324325e-05,
      "loss": 0.5608,
      "num_input_tokens_seen": 3464280,
      "step": 5295
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 1.0333082675933838,
      "learning_rate": 4.977031946381694e-05,
      "loss": 0.6381,
      "num_input_tokens_seen": 3467320,
      "step": 5300
    },
    {
      "epoch": 2.780398322851153,
      "grad_norm": 0.4984983205795288,
      "learning_rate": 4.9768770492722394e-05,
      "loss": 0.4933,
      "num_input_tokens_seen": 3470520,
      "step": 5305
    },
    {
      "epoch": 2.7830188679245285,
      "grad_norm": 0.5077530145645142,
      "learning_rate": 4.976721634028363e-05,
      "loss": 0.5106,
      "num_input_tokens_seen": 3474008,
      "step": 5310
    },
    {
      "epoch": 2.7856394129979036,
      "grad_norm": 0.9736875295639038,
      "learning_rate": 4.976565700682578e-05,
      "loss": 0.4569,
      "num_input_tokens_seen": 3477208,
      "step": 5315
    },
    {
      "epoch": 2.788259958071279,
      "grad_norm": 0.430254727602005,
      "learning_rate": 4.976409249267502e-05,
      "loss": 0.539,
      "num_input_tokens_seen": 3480088,
      "step": 5320
    },
    {
      "epoch": 2.790880503144654,
      "grad_norm": 0.24348321557044983,
      "learning_rate": 4.976252279815863e-05,
      "loss": 0.4918,
      "num_input_tokens_seen": 3484504,
      "step": 5325
    },
    {
      "epoch": 2.7935010482180296,
      "grad_norm": 0.4604262709617615,
      "learning_rate": 4.976094792360498e-05,
      "loss": 0.4458,
      "num_input_tokens_seen": 3487000,
      "step": 5330
    },
    {
      "epoch": 2.7961215932914047,
      "grad_norm": 0.3999089002609253,
      "learning_rate": 4.975936786934351e-05,
      "loss": 0.463,
      "num_input_tokens_seen": 3490712,
      "step": 5335
    },
    {
      "epoch": 2.79874213836478,
      "grad_norm": 0.34745535254478455,
      "learning_rate": 4.975778263570475e-05,
      "loss": 0.5544,
      "num_input_tokens_seen": 3494296,
      "step": 5340
    },
    {
      "epoch": 2.801362683438155,
      "grad_norm": 0.35246703028678894,
      "learning_rate": 4.975619222302032e-05,
      "loss": 0.4418,
      "num_input_tokens_seen": 3498232,
      "step": 5345
    },
    {
      "epoch": 2.8039832285115303,
      "grad_norm": 0.6888269186019897,
      "learning_rate": 4.975459663162291e-05,
      "loss": 0.4733,
      "num_input_tokens_seen": 3501080,
      "step": 5350
    },
    {
      "epoch": 2.8066037735849054,
      "grad_norm": 0.5325538516044617,
      "learning_rate": 4.9752995861846296e-05,
      "loss": 0.4214,
      "num_input_tokens_seen": 3504024,
      "step": 5355
    },
    {
      "epoch": 2.809224318658281,
      "grad_norm": 0.4960017204284668,
      "learning_rate": 4.975138991402535e-05,
      "loss": 0.4008,
      "num_input_tokens_seen": 3507256,
      "step": 5360
    },
    {
      "epoch": 2.811844863731656,
      "grad_norm": 0.29223203659057617,
      "learning_rate": 4.974977878849601e-05,
      "loss": 0.5438,
      "num_input_tokens_seen": 3510392,
      "step": 5365
    },
    {
      "epoch": 2.8144654088050314,
      "grad_norm": 0.4444985091686249,
      "learning_rate": 4.9748162485595303e-05,
      "loss": 0.5446,
      "num_input_tokens_seen": 3513272,
      "step": 5370
    },
    {
      "epoch": 2.8170859538784065,
      "grad_norm": 0.30896469950675964,
      "learning_rate": 4.974654100566135e-05,
      "loss": 0.5406,
      "num_input_tokens_seen": 3516312,
      "step": 5375
    },
    {
      "epoch": 2.819706498951782,
      "grad_norm": 0.6199353337287903,
      "learning_rate": 4.974491434903335e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 3519032,
      "step": 5380
    },
    {
      "epoch": 2.8223270440251573,
      "grad_norm": 0.5623209476470947,
      "learning_rate": 4.9743282516051556e-05,
      "loss": 0.3793,
      "num_input_tokens_seen": 3521720,
      "step": 5385
    },
    {
      "epoch": 2.8249475890985325,
      "grad_norm": 0.6982384920120239,
      "learning_rate": 4.974164550705736e-05,
      "loss": 0.4688,
      "num_input_tokens_seen": 3524888,
      "step": 5390
    },
    {
      "epoch": 2.8275681341719077,
      "grad_norm": 0.3197042942047119,
      "learning_rate": 4.974000332239319e-05,
      "loss": 0.4225,
      "num_input_tokens_seen": 3529688,
      "step": 5395
    },
    {
      "epoch": 2.830188679245283,
      "grad_norm": 0.4181956350803375,
      "learning_rate": 4.973835596240257e-05,
      "loss": 0.5128,
      "num_input_tokens_seen": 3532440,
      "step": 5400
    },
    {
      "epoch": 2.832809224318658,
      "grad_norm": 0.5415747165679932,
      "learning_rate": 4.9736703427430114e-05,
      "loss": 0.5056,
      "num_input_tokens_seen": 3535288,
      "step": 5405
    },
    {
      "epoch": 2.8354297693920336,
      "grad_norm": 0.7039856314659119,
      "learning_rate": 4.9735045717821514e-05,
      "loss": 0.5324,
      "num_input_tokens_seen": 3538840,
      "step": 5410
    },
    {
      "epoch": 2.838050314465409,
      "grad_norm": 0.3358006179332733,
      "learning_rate": 4.973338283392354e-05,
      "loss": 0.3793,
      "num_input_tokens_seen": 3541880,
      "step": 5415
    },
    {
      "epoch": 2.840670859538784,
      "grad_norm": 0.4729805886745453,
      "learning_rate": 4.9731714776084046e-05,
      "loss": 0.5319,
      "num_input_tokens_seen": 3545208,
      "step": 5420
    },
    {
      "epoch": 2.8432914046121596,
      "grad_norm": 0.7078300714492798,
      "learning_rate": 4.973004154465198e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 3548120,
      "step": 5425
    },
    {
      "epoch": 2.8459119496855347,
      "grad_norm": 0.5970824956893921,
      "learning_rate": 4.972836313997736e-05,
      "loss": 0.5252,
      "num_input_tokens_seen": 3551512,
      "step": 5430
    },
    {
      "epoch": 2.84853249475891,
      "grad_norm": 0.5871820449829102,
      "learning_rate": 4.9726679562411285e-05,
      "loss": 0.3983,
      "num_input_tokens_seen": 3554520,
      "step": 5435
    },
    {
      "epoch": 2.851153039832285,
      "grad_norm": 0.3718087077140808,
      "learning_rate": 4.972499081230595e-05,
      "loss": 0.5839,
      "num_input_tokens_seen": 3557208,
      "step": 5440
    },
    {
      "epoch": 2.8537735849056602,
      "grad_norm": 0.8063954710960388,
      "learning_rate": 4.9723296890014605e-05,
      "loss": 0.6565,
      "num_input_tokens_seen": 3560600,
      "step": 5445
    },
    {
      "epoch": 2.8563941299790354,
      "grad_norm": 0.3181860148906708,
      "learning_rate": 4.9721597795891605e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 3565752,
      "step": 5450
    },
    {
      "epoch": 2.859014675052411,
      "grad_norm": 0.2950049042701721,
      "learning_rate": 4.9719893530292396e-05,
      "loss": 0.4073,
      "num_input_tokens_seen": 3569400,
      "step": 5455
    },
    {
      "epoch": 2.861635220125786,
      "grad_norm": 1.2065622806549072,
      "learning_rate": 4.9718184093573475e-05,
      "loss": 0.4845,
      "num_input_tokens_seen": 3572088,
      "step": 5460
    },
    {
      "epoch": 2.8642557651991614,
      "grad_norm": 0.28243765234947205,
      "learning_rate": 4.9716469486092435e-05,
      "loss": 0.5887,
      "num_input_tokens_seen": 3575032,
      "step": 5465
    },
    {
      "epoch": 2.8668763102725365,
      "grad_norm": 0.5252074599266052,
      "learning_rate": 4.971474970820797e-05,
      "loss": 0.4832,
      "num_input_tokens_seen": 3578232,
      "step": 5470
    },
    {
      "epoch": 2.869496855345912,
      "grad_norm": 0.39311549067497253,
      "learning_rate": 4.971302476027982e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 3581304,
      "step": 5475
    },
    {
      "epoch": 2.8721174004192873,
      "grad_norm": 0.4814019203186035,
      "learning_rate": 4.971129464266884e-05,
      "loss": 0.4025,
      "num_input_tokens_seen": 3584664,
      "step": 5480
    },
    {
      "epoch": 2.8747379454926625,
      "grad_norm": 0.5541597008705139,
      "learning_rate": 4.970955935573693e-05,
      "loss": 0.5865,
      "num_input_tokens_seen": 3587544,
      "step": 5485
    },
    {
      "epoch": 2.8773584905660377,
      "grad_norm": 0.5916731357574463,
      "learning_rate": 4.9707818899847106e-05,
      "loss": 0.5332,
      "num_input_tokens_seen": 3590360,
      "step": 5490
    },
    {
      "epoch": 2.879979035639413,
      "grad_norm": 0.4082089066505432,
      "learning_rate": 4.970607327536346e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 3597208,
      "step": 5495
    },
    {
      "epoch": 2.882599580712788,
      "grad_norm": 0.6670106053352356,
      "learning_rate": 4.970432248265114e-05,
      "loss": 0.4045,
      "num_input_tokens_seen": 3600568,
      "step": 5500
    },
    {
      "epoch": 2.8852201257861636,
      "grad_norm": 1.1015520095825195,
      "learning_rate": 4.97025665220764e-05,
      "loss": 0.5214,
      "num_input_tokens_seen": 3603480,
      "step": 5505
    },
    {
      "epoch": 2.8878406708595388,
      "grad_norm": 0.44291847944259644,
      "learning_rate": 4.970080539400656e-05,
      "loss": 0.5617,
      "num_input_tokens_seen": 3606360,
      "step": 5510
    },
    {
      "epoch": 2.890461215932914,
      "grad_norm": 0.85169517993927,
      "learning_rate": 4.969903909881004e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 3608856,
      "step": 5515
    },
    {
      "epoch": 2.8930817610062896,
      "grad_norm": 0.3939032256603241,
      "learning_rate": 4.9697267636856317e-05,
      "loss": 0.4064,
      "num_input_tokens_seen": 3611384,
      "step": 5520
    },
    {
      "epoch": 2.8957023060796647,
      "grad_norm": 0.47770243883132935,
      "learning_rate": 4.969549100851597e-05,
      "loss": 0.7154,
      "num_input_tokens_seen": 3614488,
      "step": 5525
    },
    {
      "epoch": 2.89832285115304,
      "grad_norm": 0.49155503511428833,
      "learning_rate": 4.9693709214160645e-05,
      "loss": 0.4849,
      "num_input_tokens_seen": 3617080,
      "step": 5530
    },
    {
      "epoch": 2.900943396226415,
      "grad_norm": 0.896949052810669,
      "learning_rate": 4.969192225416307e-05,
      "loss": 0.6342,
      "num_input_tokens_seen": 3620728,
      "step": 5535
    },
    {
      "epoch": 2.9035639412997902,
      "grad_norm": 0.40260180830955505,
      "learning_rate": 4.969013012889706e-05,
      "loss": 0.55,
      "num_input_tokens_seen": 3624184,
      "step": 5540
    },
    {
      "epoch": 2.9061844863731654,
      "grad_norm": 0.25139597058296204,
      "learning_rate": 4.9688332838737504e-05,
      "loss": 0.4091,
      "num_input_tokens_seen": 3628856,
      "step": 5545
    },
    {
      "epoch": 2.908805031446541,
      "grad_norm": 0.3392549753189087,
      "learning_rate": 4.9686530384060386e-05,
      "loss": 0.5844,
      "num_input_tokens_seen": 3631800,
      "step": 5550
    },
    {
      "epoch": 2.911425576519916,
      "grad_norm": 1.1940767765045166,
      "learning_rate": 4.968472276524275e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 3634680,
      "step": 5555
    },
    {
      "epoch": 2.9140461215932913,
      "grad_norm": 0.3890454173088074,
      "learning_rate": 4.9682909982662725e-05,
      "loss": 0.4257,
      "num_input_tokens_seen": 3638264,
      "step": 5560
    },
    {
      "epoch": 2.9166666666666665,
      "grad_norm": 0.7284528017044067,
      "learning_rate": 4.968109203669953e-05,
      "loss": 0.4472,
      "num_input_tokens_seen": 3641240,
      "step": 5565
    },
    {
      "epoch": 2.919287211740042,
      "grad_norm": 0.4278191030025482,
      "learning_rate": 4.967926892773346e-05,
      "loss": 0.5792,
      "num_input_tokens_seen": 3643640,
      "step": 5570
    },
    {
      "epoch": 2.9219077568134173,
      "grad_norm": 0.2865220606327057,
      "learning_rate": 4.967744065614588e-05,
      "loss": 0.5044,
      "num_input_tokens_seen": 3648088,
      "step": 5575
    },
    {
      "epoch": 2.9245283018867925,
      "grad_norm": 0.31842222809791565,
      "learning_rate": 4.9675607222319255e-05,
      "loss": 0.4534,
      "num_input_tokens_seen": 3651352,
      "step": 5580
    },
    {
      "epoch": 2.9271488469601676,
      "grad_norm": 0.32005366683006287,
      "learning_rate": 4.967376862663711e-05,
      "loss": 0.5245,
      "num_input_tokens_seen": 3655288,
      "step": 5585
    },
    {
      "epoch": 2.929769392033543,
      "grad_norm": 0.5055784583091736,
      "learning_rate": 4.9671924869484064e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 3657624,
      "step": 5590
    },
    {
      "epoch": 2.932389937106918,
      "grad_norm": 0.4334794878959656,
      "learning_rate": 4.967007595124581e-05,
      "loss": 0.3781,
      "num_input_tokens_seen": 3660472,
      "step": 5595
    },
    {
      "epoch": 2.9350104821802936,
      "grad_norm": 0.348095178604126,
      "learning_rate": 4.966822187230911e-05,
      "loss": 0.5862,
      "num_input_tokens_seen": 3663992,
      "step": 5600
    },
    {
      "epoch": 2.9376310272536688,
      "grad_norm": 0.4967486262321472,
      "learning_rate": 4.966636263306183e-05,
      "loss": 0.5364,
      "num_input_tokens_seen": 3666424,
      "step": 5605
    },
    {
      "epoch": 2.940251572327044,
      "grad_norm": 0.4938605725765228,
      "learning_rate": 4.9664498233892895e-05,
      "loss": 0.5561,
      "num_input_tokens_seen": 3669400,
      "step": 5610
    },
    {
      "epoch": 2.9428721174004195,
      "grad_norm": 0.48534321784973145,
      "learning_rate": 4.966262867519232e-05,
      "loss": 0.6394,
      "num_input_tokens_seen": 3672440,
      "step": 5615
    },
    {
      "epoch": 2.9454926624737947,
      "grad_norm": 0.8627610802650452,
      "learning_rate": 4.966075395735119e-05,
      "loss": 0.3864,
      "num_input_tokens_seen": 3675704,
      "step": 5620
    },
    {
      "epoch": 2.94811320754717,
      "grad_norm": 0.5800425410270691,
      "learning_rate": 4.965887408076167e-05,
      "loss": 0.6551,
      "num_input_tokens_seen": 3679160,
      "step": 5625
    },
    {
      "epoch": 2.950733752620545,
      "grad_norm": 0.34015515446662903,
      "learning_rate": 4.965698904581703e-05,
      "loss": 0.4738,
      "num_input_tokens_seen": 3682328,
      "step": 5630
    },
    {
      "epoch": 2.95335429769392,
      "grad_norm": 0.5139045715332031,
      "learning_rate": 4.965509885291157e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 3685240,
      "step": 5635
    },
    {
      "epoch": 2.9559748427672954,
      "grad_norm": 0.719420313835144,
      "learning_rate": 4.9653203502440716e-05,
      "loss": 0.6449,
      "num_input_tokens_seen": 3687960,
      "step": 5640
    },
    {
      "epoch": 2.958595387840671,
      "grad_norm": 0.5375469326972961,
      "learning_rate": 4.965130299480095e-05,
      "loss": 0.4147,
      "num_input_tokens_seen": 3691704,
      "step": 5645
    },
    {
      "epoch": 2.961215932914046,
      "grad_norm": 0.503835141658783,
      "learning_rate": 4.9649397330389834e-05,
      "loss": 0.5476,
      "num_input_tokens_seen": 3695416,
      "step": 5650
    },
    {
      "epoch": 2.9638364779874213,
      "grad_norm": 0.6027949452400208,
      "learning_rate": 4.964748650960601e-05,
      "loss": 0.4738,
      "num_input_tokens_seen": 3698168,
      "step": 5655
    },
    {
      "epoch": 2.9664570230607965,
      "grad_norm": 0.6560516953468323,
      "learning_rate": 4.96455705328492e-05,
      "loss": 0.6347,
      "num_input_tokens_seen": 3700984,
      "step": 5660
    },
    {
      "epoch": 2.969077568134172,
      "grad_norm": 0.5196347236633301,
      "learning_rate": 4.964364940052021e-05,
      "loss": 0.6849,
      "num_input_tokens_seen": 3704472,
      "step": 5665
    },
    {
      "epoch": 2.9716981132075473,
      "grad_norm": 0.40681391954421997,
      "learning_rate": 4.964172311302091e-05,
      "loss": 0.4013,
      "num_input_tokens_seen": 3707960,
      "step": 5670
    },
    {
      "epoch": 2.9743186582809225,
      "grad_norm": 1.4843173027038574,
      "learning_rate": 4.9639791670754265e-05,
      "loss": 0.4852,
      "num_input_tokens_seen": 3710520,
      "step": 5675
    },
    {
      "epoch": 2.9769392033542976,
      "grad_norm": 0.5877829790115356,
      "learning_rate": 4.9637855074124305e-05,
      "loss": 0.5145,
      "num_input_tokens_seen": 3713688,
      "step": 5680
    },
    {
      "epoch": 2.979559748427673,
      "grad_norm": 0.4918762743473053,
      "learning_rate": 4.963591332353615e-05,
      "loss": 0.4629,
      "num_input_tokens_seen": 3717304,
      "step": 5685
    },
    {
      "epoch": 2.982180293501048,
      "grad_norm": 0.3591843247413635,
      "learning_rate": 4.9633966419396e-05,
      "loss": 0.5652,
      "num_input_tokens_seen": 3720888,
      "step": 5690
    },
    {
      "epoch": 2.9848008385744236,
      "grad_norm": 0.43845635652542114,
      "learning_rate": 4.96320143621111e-05,
      "loss": 0.5563,
      "num_input_tokens_seen": 3724888,
      "step": 5695
    },
    {
      "epoch": 2.9874213836477987,
      "grad_norm": 0.6735464930534363,
      "learning_rate": 4.963005715208981e-05,
      "loss": 0.4099,
      "num_input_tokens_seen": 3727288,
      "step": 5700
    },
    {
      "epoch": 2.990041928721174,
      "grad_norm": 0.7672467231750488,
      "learning_rate": 4.962809478974157e-05,
      "loss": 0.579,
      "num_input_tokens_seen": 3730328,
      "step": 5705
    },
    {
      "epoch": 2.9926624737945495,
      "grad_norm": 0.5671082735061646,
      "learning_rate": 4.962612727547686e-05,
      "loss": 0.4399,
      "num_input_tokens_seen": 3733912,
      "step": 5710
    },
    {
      "epoch": 2.9952830188679247,
      "grad_norm": 0.36546972393989563,
      "learning_rate": 4.962415460970728e-05,
      "loss": 0.4544,
      "num_input_tokens_seen": 3737464,
      "step": 5715
    },
    {
      "epoch": 2.9979035639413,
      "grad_norm": 0.38845735788345337,
      "learning_rate": 4.962217679284548e-05,
      "loss": 0.4395,
      "num_input_tokens_seen": 3740152,
      "step": 5720
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5037400722503662,
      "eval_runtime": 9.7762,
      "eval_samples_per_second": 86.741,
      "eval_steps_per_second": 21.685,
      "num_input_tokens_seen": 3742552,
      "step": 5724
    },
    {
      "epoch": 3.000524109014675,
      "grad_norm": 0.7452320456504822,
      "learning_rate": 4.962019382530521e-05,
      "loss": 0.4536,
      "num_input_tokens_seen": 3742968,
      "step": 5725
    },
    {
      "epoch": 3.00314465408805,
      "grad_norm": 0.34670600295066833,
      "learning_rate": 4.9618205707501256e-05,
      "loss": 0.4994,
      "num_input_tokens_seen": 3745880,
      "step": 5730
    },
    {
      "epoch": 3.0057651991614254,
      "grad_norm": 0.3075350821018219,
      "learning_rate": 4.961621243984954e-05,
      "loss": 0.5627,
      "num_input_tokens_seen": 3749464,
      "step": 5735
    },
    {
      "epoch": 3.008385744234801,
      "grad_norm": 0.880711019039154,
      "learning_rate": 4.9614214022767014e-05,
      "loss": 0.4372,
      "num_input_tokens_seen": 3752248,
      "step": 5740
    },
    {
      "epoch": 3.011006289308176,
      "grad_norm": 0.6339763402938843,
      "learning_rate": 4.9612210456671735e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 3755480,
      "step": 5745
    },
    {
      "epoch": 3.0136268343815513,
      "grad_norm": 0.3456425368785858,
      "learning_rate": 4.961020174198281e-05,
      "loss": 0.4432,
      "num_input_tokens_seen": 3758040,
      "step": 5750
    },
    {
      "epoch": 3.0162473794549265,
      "grad_norm": 1.01325523853302,
      "learning_rate": 4.9608187879120454e-05,
      "loss": 0.4764,
      "num_input_tokens_seen": 3761368,
      "step": 5755
    },
    {
      "epoch": 3.018867924528302,
      "grad_norm": 0.45443999767303467,
      "learning_rate": 4.9606168868505936e-05,
      "loss": 0.4953,
      "num_input_tokens_seen": 3764824,
      "step": 5760
    },
    {
      "epoch": 3.0214884696016773,
      "grad_norm": 0.3690003752708435,
      "learning_rate": 4.9604144710561616e-05,
      "loss": 0.4105,
      "num_input_tokens_seen": 3768056,
      "step": 5765
    },
    {
      "epoch": 3.0241090146750524,
      "grad_norm": 0.5519070029258728,
      "learning_rate": 4.960211540571092e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 3770168,
      "step": 5770
    },
    {
      "epoch": 3.0267295597484276,
      "grad_norm": 0.29267260432243347,
      "learning_rate": 4.960008095437836e-05,
      "loss": 0.5091,
      "num_input_tokens_seen": 3773112,
      "step": 5775
    },
    {
      "epoch": 3.029350104821803,
      "grad_norm": 0.3457060158252716,
      "learning_rate": 4.959804135698951e-05,
      "loss": 0.5413,
      "num_input_tokens_seen": 3776568,
      "step": 5780
    },
    {
      "epoch": 3.0319706498951784,
      "grad_norm": 0.28515487909317017,
      "learning_rate": 4.959599661397104e-05,
      "loss": 0.6702,
      "num_input_tokens_seen": 3780920,
      "step": 5785
    },
    {
      "epoch": 3.0345911949685536,
      "grad_norm": 0.4212001860141754,
      "learning_rate": 4.959394672575068e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 3783960,
      "step": 5790
    },
    {
      "epoch": 3.0372117400419287,
      "grad_norm": 0.5002181529998779,
      "learning_rate": 4.959189169275725e-05,
      "loss": 0.4614,
      "num_input_tokens_seen": 3786968,
      "step": 5795
    },
    {
      "epoch": 3.039832285115304,
      "grad_norm": 0.3523862063884735,
      "learning_rate": 4.958983151542065e-05,
      "loss": 0.5564,
      "num_input_tokens_seen": 3790808,
      "step": 5800
    },
    {
      "epoch": 3.042452830188679,
      "grad_norm": 0.3420685827732086,
      "learning_rate": 4.958776619417182e-05,
      "loss": 0.4891,
      "num_input_tokens_seen": 3794264,
      "step": 5805
    },
    {
      "epoch": 3.0450733752620547,
      "grad_norm": 0.3888975977897644,
      "learning_rate": 4.9585695729442814e-05,
      "loss": 0.5487,
      "num_input_tokens_seen": 3797496,
      "step": 5810
    },
    {
      "epoch": 3.04769392033543,
      "grad_norm": 0.5420519709587097,
      "learning_rate": 4.9583620121666754e-05,
      "loss": 0.6303,
      "num_input_tokens_seen": 3800888,
      "step": 5815
    },
    {
      "epoch": 3.050314465408805,
      "grad_norm": 0.3483212888240814,
      "learning_rate": 4.958153937127783e-05,
      "loss": 0.4401,
      "num_input_tokens_seen": 3803512,
      "step": 5820
    },
    {
      "epoch": 3.05293501048218,
      "grad_norm": 0.8193895220756531,
      "learning_rate": 4.95794534787113e-05,
      "loss": 0.5872,
      "num_input_tokens_seen": 3806552,
      "step": 5825
    },
    {
      "epoch": 3.0555555555555554,
      "grad_norm": 0.2388603538274765,
      "learning_rate": 4.957736244440353e-05,
      "loss": 0.5616,
      "num_input_tokens_seen": 3810648,
      "step": 5830
    },
    {
      "epoch": 3.058176100628931,
      "grad_norm": 0.32299336791038513,
      "learning_rate": 4.957526626879193e-05,
      "loss": 0.4085,
      "num_input_tokens_seen": 3813976,
      "step": 5835
    },
    {
      "epoch": 3.060796645702306,
      "grad_norm": 1.25869619846344,
      "learning_rate": 4.957316495231499e-05,
      "loss": 0.5506,
      "num_input_tokens_seen": 3816824,
      "step": 5840
    },
    {
      "epoch": 3.0634171907756813,
      "grad_norm": 0.2599155306816101,
      "learning_rate": 4.9571058495412294e-05,
      "loss": 0.6126,
      "num_input_tokens_seen": 3820376,
      "step": 5845
    },
    {
      "epoch": 3.0660377358490565,
      "grad_norm": 0.600284218788147,
      "learning_rate": 4.9568946898524485e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 3823736,
      "step": 5850
    },
    {
      "epoch": 3.068658280922432,
      "grad_norm": 0.5165765881538391,
      "learning_rate": 4.956683016209327e-05,
      "loss": 0.6417,
      "num_input_tokens_seen": 3826104,
      "step": 5855
    },
    {
      "epoch": 3.0712788259958073,
      "grad_norm": 0.40501144528388977,
      "learning_rate": 4.956470828656147e-05,
      "loss": 0.5584,
      "num_input_tokens_seen": 3828632,
      "step": 5860
    },
    {
      "epoch": 3.0738993710691824,
      "grad_norm": 0.32664233446121216,
      "learning_rate": 4.956258127237294e-05,
      "loss": 0.3347,
      "num_input_tokens_seen": 3831576,
      "step": 5865
    },
    {
      "epoch": 3.0765199161425576,
      "grad_norm": 0.3766722083091736,
      "learning_rate": 4.956044911997263e-05,
      "loss": 0.3617,
      "num_input_tokens_seen": 3834392,
      "step": 5870
    },
    {
      "epoch": 3.0791404612159328,
      "grad_norm": 0.6080725789070129,
      "learning_rate": 4.955831182980657e-05,
      "loss": 0.5556,
      "num_input_tokens_seen": 3837144,
      "step": 5875
    },
    {
      "epoch": 3.0817610062893084,
      "grad_norm": 0.4701588451862335,
      "learning_rate": 4.955616940232185e-05,
      "loss": 0.5569,
      "num_input_tokens_seen": 3840504,
      "step": 5880
    },
    {
      "epoch": 3.0843815513626835,
      "grad_norm": 0.33324068784713745,
      "learning_rate": 4.9554021837966636e-05,
      "loss": 0.5395,
      "num_input_tokens_seen": 3844408,
      "step": 5885
    },
    {
      "epoch": 3.0870020964360587,
      "grad_norm": 0.4125520884990692,
      "learning_rate": 4.955186913719019e-05,
      "loss": 0.5374,
      "num_input_tokens_seen": 3847736,
      "step": 5890
    },
    {
      "epoch": 3.089622641509434,
      "grad_norm": 0.3695298135280609,
      "learning_rate": 4.954971130044281e-05,
      "loss": 0.5321,
      "num_input_tokens_seen": 3850360,
      "step": 5895
    },
    {
      "epoch": 3.092243186582809,
      "grad_norm": 0.37060782313346863,
      "learning_rate": 4.954754832817592e-05,
      "loss": 0.4768,
      "num_input_tokens_seen": 3853624,
      "step": 5900
    },
    {
      "epoch": 3.0948637316561847,
      "grad_norm": 0.4810468256473541,
      "learning_rate": 4.954538022084196e-05,
      "loss": 0.3917,
      "num_input_tokens_seen": 3856664,
      "step": 5905
    },
    {
      "epoch": 3.09748427672956,
      "grad_norm": 0.4957842528820038,
      "learning_rate": 4.9543206978894496e-05,
      "loss": 0.5471,
      "num_input_tokens_seen": 3859640,
      "step": 5910
    },
    {
      "epoch": 3.100104821802935,
      "grad_norm": 0.8035592436790466,
      "learning_rate": 4.954102860278813e-05,
      "loss": 0.4107,
      "num_input_tokens_seen": 3862712,
      "step": 5915
    },
    {
      "epoch": 3.10272536687631,
      "grad_norm": 0.3393804728984833,
      "learning_rate": 4.9538845092978567e-05,
      "loss": 0.4881,
      "num_input_tokens_seen": 3866328,
      "step": 5920
    },
    {
      "epoch": 3.1053459119496853,
      "grad_norm": 0.3386576771736145,
      "learning_rate": 4.953665644992256e-05,
      "loss": 0.4746,
      "num_input_tokens_seen": 3869592,
      "step": 5925
    },
    {
      "epoch": 3.107966457023061,
      "grad_norm": 0.4988095164299011,
      "learning_rate": 4.953446267407795e-05,
      "loss": 0.3941,
      "num_input_tokens_seen": 3875032,
      "step": 5930
    },
    {
      "epoch": 3.110587002096436,
      "grad_norm": 0.2535756528377533,
      "learning_rate": 4.953226376590365e-05,
      "loss": 0.5686,
      "num_input_tokens_seen": 3879288,
      "step": 5935
    },
    {
      "epoch": 3.1132075471698113,
      "grad_norm": 0.4325270354747772,
      "learning_rate": 4.953005972585965e-05,
      "loss": 0.3951,
      "num_input_tokens_seen": 3882936,
      "step": 5940
    },
    {
      "epoch": 3.1158280922431865,
      "grad_norm": 0.31084397435188293,
      "learning_rate": 4.952785055440701e-05,
      "loss": 0.5339,
      "num_input_tokens_seen": 3886296,
      "step": 5945
    },
    {
      "epoch": 3.1184486373165616,
      "grad_norm": 0.8660086393356323,
      "learning_rate": 4.9525636252007864e-05,
      "loss": 0.547,
      "num_input_tokens_seen": 3889816,
      "step": 5950
    },
    {
      "epoch": 3.1210691823899372,
      "grad_norm": 0.5637370347976685,
      "learning_rate": 4.9523416819125425e-05,
      "loss": 0.3993,
      "num_input_tokens_seen": 3892888,
      "step": 5955
    },
    {
      "epoch": 3.1236897274633124,
      "grad_norm": 0.9615580439567566,
      "learning_rate": 4.952119225622396e-05,
      "loss": 0.5678,
      "num_input_tokens_seen": 3896152,
      "step": 5960
    },
    {
      "epoch": 3.1263102725366876,
      "grad_norm": 0.40786629915237427,
      "learning_rate": 4.951896256376882e-05,
      "loss": 0.3533,
      "num_input_tokens_seen": 3898776,
      "step": 5965
    },
    {
      "epoch": 3.1289308176100628,
      "grad_norm": 0.41846218705177307,
      "learning_rate": 4.9516727742226444e-05,
      "loss": 0.5251,
      "num_input_tokens_seen": 3901912,
      "step": 5970
    },
    {
      "epoch": 3.131551362683438,
      "grad_norm": 0.32646310329437256,
      "learning_rate": 4.9514487792064323e-05,
      "loss": 0.58,
      "num_input_tokens_seen": 3905688,
      "step": 5975
    },
    {
      "epoch": 3.1341719077568135,
      "grad_norm": 0.4185785949230194,
      "learning_rate": 4.951224271375104e-05,
      "loss": 0.5091,
      "num_input_tokens_seen": 3908888,
      "step": 5980
    },
    {
      "epoch": 3.1367924528301887,
      "grad_norm": 0.5953133702278137,
      "learning_rate": 4.950999250775623e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 3912280,
      "step": 5985
    },
    {
      "epoch": 3.139412997903564,
      "grad_norm": 0.5617034435272217,
      "learning_rate": 4.950773717455061e-05,
      "loss": 0.4512,
      "num_input_tokens_seen": 3915256,
      "step": 5990
    },
    {
      "epoch": 3.142033542976939,
      "grad_norm": 0.36650872230529785,
      "learning_rate": 4.950547671460596e-05,
      "loss": 0.4588,
      "num_input_tokens_seen": 3917912,
      "step": 5995
    },
    {
      "epoch": 3.1446540880503147,
      "grad_norm": 0.43795493245124817,
      "learning_rate": 4.950321112839517e-05,
      "loss": 0.5585,
      "num_input_tokens_seen": 3920824,
      "step": 6000
    },
    {
      "epoch": 3.14727463312369,
      "grad_norm": 0.6218132376670837,
      "learning_rate": 4.9500940416392156e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 3923736,
      "step": 6005
    },
    {
      "epoch": 3.149895178197065,
      "grad_norm": 0.34272903203964233,
      "learning_rate": 4.949866457907193e-05,
      "loss": 0.5384,
      "num_input_tokens_seen": 3926584,
      "step": 6010
    },
    {
      "epoch": 3.15251572327044,
      "grad_norm": 0.6102215051651001,
      "learning_rate": 4.9496383616910565e-05,
      "loss": 0.6501,
      "num_input_tokens_seen": 3929432,
      "step": 6015
    },
    {
      "epoch": 3.1551362683438153,
      "grad_norm": 0.6061221361160278,
      "learning_rate": 4.949409753038522e-05,
      "loss": 0.5332,
      "num_input_tokens_seen": 3932088,
      "step": 6020
    },
    {
      "epoch": 3.157756813417191,
      "grad_norm": 0.40089136362075806,
      "learning_rate": 4.949180631997411e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 3935384,
      "step": 6025
    },
    {
      "epoch": 3.160377358490566,
      "grad_norm": 0.38459038734436035,
      "learning_rate": 4.948950998615654e-05,
      "loss": 0.4594,
      "num_input_tokens_seen": 3938424,
      "step": 6030
    },
    {
      "epoch": 3.1629979035639413,
      "grad_norm": 0.4573156237602234,
      "learning_rate": 4.9487208529412876e-05,
      "loss": 0.5189,
      "num_input_tokens_seen": 3940984,
      "step": 6035
    },
    {
      "epoch": 3.1656184486373165,
      "grad_norm": 0.3921795189380646,
      "learning_rate": 4.9484901950224546e-05,
      "loss": 0.4833,
      "num_input_tokens_seen": 3945144,
      "step": 6040
    },
    {
      "epoch": 3.1682389937106916,
      "grad_norm": 0.24945321679115295,
      "learning_rate": 4.948259024907407e-05,
      "loss": 0.5069,
      "num_input_tokens_seen": 3948344,
      "step": 6045
    },
    {
      "epoch": 3.1708595387840672,
      "grad_norm": 0.43722420930862427,
      "learning_rate": 4.948027342644502e-05,
      "loss": 0.5539,
      "num_input_tokens_seen": 3951000,
      "step": 6050
    },
    {
      "epoch": 3.1734800838574424,
      "grad_norm": 0.5655446648597717,
      "learning_rate": 4.947795148282206e-05,
      "loss": 0.4749,
      "num_input_tokens_seen": 3954584,
      "step": 6055
    },
    {
      "epoch": 3.1761006289308176,
      "grad_norm": 0.42747440934181213,
      "learning_rate": 4.947562441869091e-05,
      "loss": 0.3559,
      "num_input_tokens_seen": 3957944,
      "step": 6060
    },
    {
      "epoch": 3.1787211740041927,
      "grad_norm": 0.3057638704776764,
      "learning_rate": 4.947329223453836e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 3961816,
      "step": 6065
    },
    {
      "epoch": 3.181341719077568,
      "grad_norm": 0.4274386465549469,
      "learning_rate": 4.947095493085229e-05,
      "loss": 0.5366,
      "num_input_tokens_seen": 3964824,
      "step": 6070
    },
    {
      "epoch": 3.1839622641509435,
      "grad_norm": 0.4257733225822449,
      "learning_rate": 4.9468612508121616e-05,
      "loss": 0.381,
      "num_input_tokens_seen": 3968056,
      "step": 6075
    },
    {
      "epoch": 3.1865828092243187,
      "grad_norm": 0.42202678322792053,
      "learning_rate": 4.9466264966836365e-05,
      "loss": 0.4774,
      "num_input_tokens_seen": 3970424,
      "step": 6080
    },
    {
      "epoch": 3.189203354297694,
      "grad_norm": 0.7713318467140198,
      "learning_rate": 4.94639123074876e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 3973208,
      "step": 6085
    },
    {
      "epoch": 3.191823899371069,
      "grad_norm": 0.437461256980896,
      "learning_rate": 4.946155453056749e-05,
      "loss": 0.5327,
      "num_input_tokens_seen": 3976216,
      "step": 6090
    },
    {
      "epoch": 3.1944444444444446,
      "grad_norm": 0.6407696604728699,
      "learning_rate": 4.945919163656923e-05,
      "loss": 0.5511,
      "num_input_tokens_seen": 3980824,
      "step": 6095
    },
    {
      "epoch": 3.19706498951782,
      "grad_norm": 0.4399241507053375,
      "learning_rate": 4.945682362598715e-05,
      "loss": 0.5599,
      "num_input_tokens_seen": 3987480,
      "step": 6100
    },
    {
      "epoch": 3.199685534591195,
      "grad_norm": 0.41117411851882935,
      "learning_rate": 4.945445049931656e-05,
      "loss": 0.4674,
      "num_input_tokens_seen": 3991288,
      "step": 6105
    },
    {
      "epoch": 3.20230607966457,
      "grad_norm": 0.39059028029441833,
      "learning_rate": 4.945207225705393e-05,
      "loss": 0.5697,
      "num_input_tokens_seen": 3993848,
      "step": 6110
    },
    {
      "epoch": 3.2049266247379453,
      "grad_norm": 0.4444116950035095,
      "learning_rate": 4.944968889969674e-05,
      "loss": 0.5029,
      "num_input_tokens_seen": 3997848,
      "step": 6115
    },
    {
      "epoch": 3.207547169811321,
      "grad_norm": 0.5996315479278564,
      "learning_rate": 4.944730042774358e-05,
      "loss": 0.4265,
      "num_input_tokens_seen": 4001784,
      "step": 6120
    },
    {
      "epoch": 3.210167714884696,
      "grad_norm": 0.3181256055831909,
      "learning_rate": 4.944490684169408e-05,
      "loss": 0.5077,
      "num_input_tokens_seen": 4005048,
      "step": 6125
    },
    {
      "epoch": 3.2127882599580713,
      "grad_norm": 0.8743860721588135,
      "learning_rate": 4.9442508142048945e-05,
      "loss": 0.5192,
      "num_input_tokens_seen": 4007864,
      "step": 6130
    },
    {
      "epoch": 3.2154088050314464,
      "grad_norm": 0.40847182273864746,
      "learning_rate": 4.9440104329309956e-05,
      "loss": 0.5436,
      "num_input_tokens_seen": 4011832,
      "step": 6135
    },
    {
      "epoch": 3.2180293501048216,
      "grad_norm": 0.37415868043899536,
      "learning_rate": 4.943769540397998e-05,
      "loss": 0.4777,
      "num_input_tokens_seen": 4015032,
      "step": 6140
    },
    {
      "epoch": 3.220649895178197,
      "grad_norm": 0.4735053777694702,
      "learning_rate": 4.943528136656292e-05,
      "loss": 0.606,
      "num_input_tokens_seen": 4017432,
      "step": 6145
    },
    {
      "epoch": 3.2232704402515724,
      "grad_norm": 0.6722296476364136,
      "learning_rate": 4.943286221756377e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 4020216,
      "step": 6150
    },
    {
      "epoch": 3.2258909853249476,
      "grad_norm": 0.5493027567863464,
      "learning_rate": 4.9430437957488586e-05,
      "loss": 0.5535,
      "num_input_tokens_seen": 4023096,
      "step": 6155
    },
    {
      "epoch": 3.2285115303983227,
      "grad_norm": 0.9303476214408875,
      "learning_rate": 4.94280085868445e-05,
      "loss": 0.395,
      "num_input_tokens_seen": 4026712,
      "step": 6160
    },
    {
      "epoch": 3.231132075471698,
      "grad_norm": 0.8441289663314819,
      "learning_rate": 4.942557410613972e-05,
      "loss": 0.5412,
      "num_input_tokens_seen": 4030872,
      "step": 6165
    },
    {
      "epoch": 3.2337526205450735,
      "grad_norm": 0.4168168604373932,
      "learning_rate": 4.942313451588348e-05,
      "loss": 0.4506,
      "num_input_tokens_seen": 4034744,
      "step": 6170
    },
    {
      "epoch": 3.2363731656184487,
      "grad_norm": 0.5498723387718201,
      "learning_rate": 4.942068981658614e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 4037752,
      "step": 6175
    },
    {
      "epoch": 3.238993710691824,
      "grad_norm": 1.0890499353408813,
      "learning_rate": 4.9418240008759097e-05,
      "loss": 0.4415,
      "num_input_tokens_seen": 4041048,
      "step": 6180
    },
    {
      "epoch": 3.241614255765199,
      "grad_norm": 0.6905120015144348,
      "learning_rate": 4.9415785092914826e-05,
      "loss": 0.583,
      "num_input_tokens_seen": 4043832,
      "step": 6185
    },
    {
      "epoch": 3.2442348008385746,
      "grad_norm": 0.494446337223053,
      "learning_rate": 4.941332506956686e-05,
      "loss": 0.4298,
      "num_input_tokens_seen": 4047512,
      "step": 6190
    },
    {
      "epoch": 3.24685534591195,
      "grad_norm": 0.39334502816200256,
      "learning_rate": 4.941085993922981e-05,
      "loss": 0.497,
      "num_input_tokens_seen": 4050584,
      "step": 6195
    },
    {
      "epoch": 3.249475890985325,
      "grad_norm": 0.47529831528663635,
      "learning_rate": 4.9408389702419354e-05,
      "loss": 0.3939,
      "num_input_tokens_seen": 4053752,
      "step": 6200
    },
    {
      "epoch": 3.2520964360587,
      "grad_norm": 0.7832316160202026,
      "learning_rate": 4.940591435965224e-05,
      "loss": 0.511,
      "num_input_tokens_seen": 4056824,
      "step": 6205
    },
    {
      "epoch": 3.2547169811320753,
      "grad_norm": 1.0177167654037476,
      "learning_rate": 4.9403433911446285e-05,
      "loss": 0.6433,
      "num_input_tokens_seen": 4059864,
      "step": 6210
    },
    {
      "epoch": 3.257337526205451,
      "grad_norm": 0.6382784843444824,
      "learning_rate": 4.9400948358320356e-05,
      "loss": 0.4699,
      "num_input_tokens_seen": 4063416,
      "step": 6215
    },
    {
      "epoch": 3.259958071278826,
      "grad_norm": 0.3250112235546112,
      "learning_rate": 4.939845770079442e-05,
      "loss": 0.4225,
      "num_input_tokens_seen": 4066264,
      "step": 6220
    },
    {
      "epoch": 3.2625786163522013,
      "grad_norm": 0.4587024748325348,
      "learning_rate": 4.9395961939389475e-05,
      "loss": 0.5196,
      "num_input_tokens_seen": 4068664,
      "step": 6225
    },
    {
      "epoch": 3.2651991614255764,
      "grad_norm": 0.5107795596122742,
      "learning_rate": 4.939346107462762e-05,
      "loss": 0.6998,
      "num_input_tokens_seen": 4072088,
      "step": 6230
    },
    {
      "epoch": 3.2678197064989516,
      "grad_norm": 1.0975347757339478,
      "learning_rate": 4.9390955107032e-05,
      "loss": 0.6679,
      "num_input_tokens_seen": 4075064,
      "step": 6235
    },
    {
      "epoch": 3.270440251572327,
      "grad_norm": 0.551445722579956,
      "learning_rate": 4.938844403712685e-05,
      "loss": 0.521,
      "num_input_tokens_seen": 4077752,
      "step": 6240
    },
    {
      "epoch": 3.2730607966457024,
      "grad_norm": 0.4975104331970215,
      "learning_rate": 4.938592786543743e-05,
      "loss": 0.5181,
      "num_input_tokens_seen": 4080472,
      "step": 6245
    },
    {
      "epoch": 3.2756813417190775,
      "grad_norm": 0.5820958614349365,
      "learning_rate": 4.938340659249012e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 4083384,
      "step": 6250
    },
    {
      "epoch": 3.2783018867924527,
      "grad_norm": 0.36260315775871277,
      "learning_rate": 4.938088021881233e-05,
      "loss": 0.5343,
      "num_input_tokens_seen": 4086488,
      "step": 6255
    },
    {
      "epoch": 3.280922431865828,
      "grad_norm": 0.4233722984790802,
      "learning_rate": 4.937834874493255e-05,
      "loss": 0.3484,
      "num_input_tokens_seen": 4089464,
      "step": 6260
    },
    {
      "epoch": 3.2835429769392035,
      "grad_norm": 0.5350418090820312,
      "learning_rate": 4.9375812171380334e-05,
      "loss": 0.5469,
      "num_input_tokens_seen": 4092088,
      "step": 6265
    },
    {
      "epoch": 3.2861635220125787,
      "grad_norm": 0.5555964112281799,
      "learning_rate": 4.9373270498686306e-05,
      "loss": 0.4819,
      "num_input_tokens_seen": 4095544,
      "step": 6270
    },
    {
      "epoch": 3.288784067085954,
      "grad_norm": 0.26828110218048096,
      "learning_rate": 4.937072372738216e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 4098264,
      "step": 6275
    },
    {
      "epoch": 3.291404612159329,
      "grad_norm": 0.44629430770874023,
      "learning_rate": 4.9368171858000635e-05,
      "loss": 0.4603,
      "num_input_tokens_seen": 4100792,
      "step": 6280
    },
    {
      "epoch": 3.2940251572327046,
      "grad_norm": 0.35368624329566956,
      "learning_rate": 4.9365614891075576e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 4103896,
      "step": 6285
    },
    {
      "epoch": 3.29664570230608,
      "grad_norm": 0.3815270662307739,
      "learning_rate": 4.936305282714185e-05,
      "loss": 0.3859,
      "num_input_tokens_seen": 4106968,
      "step": 6290
    },
    {
      "epoch": 3.299266247379455,
      "grad_norm": 0.41380882263183594,
      "learning_rate": 4.936048566673542e-05,
      "loss": 0.5584,
      "num_input_tokens_seen": 4110040,
      "step": 6295
    },
    {
      "epoch": 3.30188679245283,
      "grad_norm": 0.4148041903972626,
      "learning_rate": 4.935791341039332e-05,
      "loss": 0.4478,
      "num_input_tokens_seen": 4112664,
      "step": 6300
    },
    {
      "epoch": 3.3045073375262053,
      "grad_norm": 0.8245702385902405,
      "learning_rate": 4.935533605865361e-05,
      "loss": 0.4943,
      "num_input_tokens_seen": 4115832,
      "step": 6305
    },
    {
      "epoch": 3.307127882599581,
      "grad_norm": 0.5183990597724915,
      "learning_rate": 4.935275361205547e-05,
      "loss": 0.4661,
      "num_input_tokens_seen": 4119160,
      "step": 6310
    },
    {
      "epoch": 3.309748427672956,
      "grad_norm": 0.6890310645103455,
      "learning_rate": 4.93501660711391e-05,
      "loss": 0.8488,
      "num_input_tokens_seen": 4122520,
      "step": 6315
    },
    {
      "epoch": 3.3123689727463312,
      "grad_norm": 0.6095846891403198,
      "learning_rate": 4.934757343644578e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 4124888,
      "step": 6320
    },
    {
      "epoch": 3.3149895178197064,
      "grad_norm": 0.44577449560165405,
      "learning_rate": 4.934497570851788e-05,
      "loss": 0.5301,
      "num_input_tokens_seen": 4127832,
      "step": 6325
    },
    {
      "epoch": 3.3176100628930816,
      "grad_norm": 0.5070707201957703,
      "learning_rate": 4.9342372887898804e-05,
      "loss": 0.6176,
      "num_input_tokens_seen": 4130232,
      "step": 6330
    },
    {
      "epoch": 3.320230607966457,
      "grad_norm": 0.36468571424484253,
      "learning_rate": 4.9339764975133026e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 4133080,
      "step": 6335
    },
    {
      "epoch": 3.3228511530398324,
      "grad_norm": 0.2590022385120392,
      "learning_rate": 4.93371519707661e-05,
      "loss": 0.4335,
      "num_input_tokens_seen": 4136536,
      "step": 6340
    },
    {
      "epoch": 3.3254716981132075,
      "grad_norm": 0.32090556621551514,
      "learning_rate": 4.933453387534464e-05,
      "loss": 0.4455,
      "num_input_tokens_seen": 4139608,
      "step": 6345
    },
    {
      "epoch": 3.3280922431865827,
      "grad_norm": 0.7564888000488281,
      "learning_rate": 4.933191068941631e-05,
      "loss": 0.6084,
      "num_input_tokens_seen": 4142200,
      "step": 6350
    },
    {
      "epoch": 3.330712788259958,
      "grad_norm": 0.6555405259132385,
      "learning_rate": 4.932928241352987e-05,
      "loss": 0.83,
      "num_input_tokens_seen": 4146200,
      "step": 6355
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 0.3914836645126343,
      "learning_rate": 4.9326649048235095e-05,
      "loss": 0.4113,
      "num_input_tokens_seen": 4149464,
      "step": 6360
    },
    {
      "epoch": 3.3359538784067087,
      "grad_norm": 0.2929394245147705,
      "learning_rate": 4.932401059408287e-05,
      "loss": 0.5624,
      "num_input_tokens_seen": 4151992,
      "step": 6365
    },
    {
      "epoch": 3.338574423480084,
      "grad_norm": 0.35673075914382935,
      "learning_rate": 4.932136705162514e-05,
      "loss": 0.4974,
      "num_input_tokens_seen": 4155576,
      "step": 6370
    },
    {
      "epoch": 3.341194968553459,
      "grad_norm": 0.5364058017730713,
      "learning_rate": 4.93187184214149e-05,
      "loss": 0.6032,
      "num_input_tokens_seen": 4159672,
      "step": 6375
    },
    {
      "epoch": 3.3438155136268346,
      "grad_norm": 0.4236099123954773,
      "learning_rate": 4.9316064704006205e-05,
      "loss": 0.4452,
      "num_input_tokens_seen": 4162360,
      "step": 6380
    },
    {
      "epoch": 3.3464360587002098,
      "grad_norm": 0.4178391993045807,
      "learning_rate": 4.931340589995419e-05,
      "loss": 0.4738,
      "num_input_tokens_seen": 4165304,
      "step": 6385
    },
    {
      "epoch": 3.349056603773585,
      "grad_norm": 0.3170464336872101,
      "learning_rate": 4.9310742009815035e-05,
      "loss": 0.4536,
      "num_input_tokens_seen": 4168728,
      "step": 6390
    },
    {
      "epoch": 3.35167714884696,
      "grad_norm": 0.6654120683670044,
      "learning_rate": 4.930807303414601e-05,
      "loss": 0.3833,
      "num_input_tokens_seen": 4171960,
      "step": 6395
    },
    {
      "epoch": 3.3542976939203353,
      "grad_norm": 0.3176206052303314,
      "learning_rate": 4.930539897350542e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 4175544,
      "step": 6400
    },
    {
      "epoch": 3.3569182389937104,
      "grad_norm": 0.32410839200019836,
      "learning_rate": 4.930271982845266e-05,
      "loss": 0.403,
      "num_input_tokens_seen": 4178264,
      "step": 6405
    },
    {
      "epoch": 3.359538784067086,
      "grad_norm": 0.515836238861084,
      "learning_rate": 4.930003559954817e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 4180984,
      "step": 6410
    },
    {
      "epoch": 3.3621593291404612,
      "grad_norm": 0.4780007004737854,
      "learning_rate": 4.929734628735346e-05,
      "loss": 0.5908,
      "num_input_tokens_seen": 4183768,
      "step": 6415
    },
    {
      "epoch": 3.3647798742138364,
      "grad_norm": 0.3714679479598999,
      "learning_rate": 4.9294651892431115e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 4187000,
      "step": 6420
    },
    {
      "epoch": 3.3674004192872116,
      "grad_norm": 0.7912814021110535,
      "learning_rate": 4.929195241534476e-05,
      "loss": 0.496,
      "num_input_tokens_seen": 4189304,
      "step": 6425
    },
    {
      "epoch": 3.370020964360587,
      "grad_norm": 0.3046709895133972,
      "learning_rate": 4.9289247856659095e-05,
      "loss": 0.4991,
      "num_input_tokens_seen": 4193080,
      "step": 6430
    },
    {
      "epoch": 3.3726415094339623,
      "grad_norm": 0.5698127746582031,
      "learning_rate": 4.9286538216939883e-05,
      "loss": 0.4976,
      "num_input_tokens_seen": 4195608,
      "step": 6435
    },
    {
      "epoch": 3.3752620545073375,
      "grad_norm": 0.6037057042121887,
      "learning_rate": 4.9283823496753944e-05,
      "loss": 0.3844,
      "num_input_tokens_seen": 4198744,
      "step": 6440
    },
    {
      "epoch": 3.3778825995807127,
      "grad_norm": 0.7402395009994507,
      "learning_rate": 4.928110369666918e-05,
      "loss": 0.7382,
      "num_input_tokens_seen": 4202680,
      "step": 6445
    },
    {
      "epoch": 3.380503144654088,
      "grad_norm": 0.6164316534996033,
      "learning_rate": 4.927837881725455e-05,
      "loss": 0.4779,
      "num_input_tokens_seen": 4205144,
      "step": 6450
    },
    {
      "epoch": 3.3831236897274635,
      "grad_norm": 0.5455995798110962,
      "learning_rate": 4.927564885908003e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 4209112,
      "step": 6455
    },
    {
      "epoch": 3.3857442348008386,
      "grad_norm": 0.4577008783817291,
      "learning_rate": 4.9272913822716735e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 4213048,
      "step": 6460
    },
    {
      "epoch": 3.388364779874214,
      "grad_norm": 0.28251218795776367,
      "learning_rate": 4.927017370873678e-05,
      "loss": 0.4389,
      "num_input_tokens_seen": 4216472,
      "step": 6465
    },
    {
      "epoch": 3.390985324947589,
      "grad_norm": 0.4376584589481354,
      "learning_rate": 4.926742851771338e-05,
      "loss": 0.4221,
      "num_input_tokens_seen": 4219672,
      "step": 6470
    },
    {
      "epoch": 3.3936058700209646,
      "grad_norm": 0.6302117705345154,
      "learning_rate": 4.926467825022079e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 4222872,
      "step": 6475
    },
    {
      "epoch": 3.3962264150943398,
      "grad_norm": 0.32186660170555115,
      "learning_rate": 4.926192290683434e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 4225528,
      "step": 6480
    },
    {
      "epoch": 3.398846960167715,
      "grad_norm": 0.5571965575218201,
      "learning_rate": 4.9259162488130404e-05,
      "loss": 0.4106,
      "num_input_tokens_seen": 4229144,
      "step": 6485
    },
    {
      "epoch": 3.40146750524109,
      "grad_norm": 0.38411393761634827,
      "learning_rate": 4.925639699468645e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 4232792,
      "step": 6490
    },
    {
      "epoch": 3.4040880503144653,
      "grad_norm": 0.32654711604118347,
      "learning_rate": 4.925362642708097e-05,
      "loss": 0.4396,
      "num_input_tokens_seen": 4236600,
      "step": 6495
    },
    {
      "epoch": 3.4067085953878404,
      "grad_norm": 0.5292193293571472,
      "learning_rate": 4.9250850785893536e-05,
      "loss": 0.6463,
      "num_input_tokens_seen": 4239896,
      "step": 6500
    },
    {
      "epoch": 3.409329140461216,
      "grad_norm": 0.31277230381965637,
      "learning_rate": 4.924807007170479e-05,
      "loss": 0.5118,
      "num_input_tokens_seen": 4243640,
      "step": 6505
    },
    {
      "epoch": 3.411949685534591,
      "grad_norm": 0.4299231469631195,
      "learning_rate": 4.924528428509643e-05,
      "loss": 0.5725,
      "num_input_tokens_seen": 4247160,
      "step": 6510
    },
    {
      "epoch": 3.4145702306079664,
      "grad_norm": 0.45518893003463745,
      "learning_rate": 4.9242493426651196e-05,
      "loss": 0.5936,
      "num_input_tokens_seen": 4251032,
      "step": 6515
    },
    {
      "epoch": 3.4171907756813416,
      "grad_norm": 0.6034185886383057,
      "learning_rate": 4.92396974969529e-05,
      "loss": 0.5594,
      "num_input_tokens_seen": 4254680,
      "step": 6520
    },
    {
      "epoch": 3.419811320754717,
      "grad_norm": 0.40931829810142517,
      "learning_rate": 4.923689649658644e-05,
      "loss": 0.4291,
      "num_input_tokens_seen": 4257720,
      "step": 6525
    },
    {
      "epoch": 3.4224318658280923,
      "grad_norm": 0.3173772394657135,
      "learning_rate": 4.923409042613774e-05,
      "loss": 0.5196,
      "num_input_tokens_seen": 4260696,
      "step": 6530
    },
    {
      "epoch": 3.4250524109014675,
      "grad_norm": 0.47671011090278625,
      "learning_rate": 4.92312792861938e-05,
      "loss": 0.4558,
      "num_input_tokens_seen": 4263416,
      "step": 6535
    },
    {
      "epoch": 3.4276729559748427,
      "grad_norm": 0.6289641857147217,
      "learning_rate": 4.922846307734268e-05,
      "loss": 0.6005,
      "num_input_tokens_seen": 4266232,
      "step": 6540
    },
    {
      "epoch": 3.430293501048218,
      "grad_norm": 0.9822584390640259,
      "learning_rate": 4.9225641800173496e-05,
      "loss": 0.4312,
      "num_input_tokens_seen": 4269400,
      "step": 6545
    },
    {
      "epoch": 3.4329140461215935,
      "grad_norm": 0.5642045736312866,
      "learning_rate": 4.922281545527643e-05,
      "loss": 0.4887,
      "num_input_tokens_seen": 4271896,
      "step": 6550
    },
    {
      "epoch": 3.4355345911949686,
      "grad_norm": 0.4885694682598114,
      "learning_rate": 4.9219984043242715e-05,
      "loss": 0.6836,
      "num_input_tokens_seen": 4275416,
      "step": 6555
    },
    {
      "epoch": 3.438155136268344,
      "grad_norm": 0.2893494665622711,
      "learning_rate": 4.921714756466466e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 4278936,
      "step": 6560
    },
    {
      "epoch": 3.440775681341719,
      "grad_norm": 0.4622831344604492,
      "learning_rate": 4.921430602013561e-05,
      "loss": 0.4918,
      "num_input_tokens_seen": 4282808,
      "step": 6565
    },
    {
      "epoch": 3.4433962264150946,
      "grad_norm": 0.46298831701278687,
      "learning_rate": 4.9211459410249995e-05,
      "loss": 0.5185,
      "num_input_tokens_seen": 4285400,
      "step": 6570
    },
    {
      "epoch": 3.4460167714884697,
      "grad_norm": 0.7406746745109558,
      "learning_rate": 4.92086077356033e-05,
      "loss": 0.4025,
      "num_input_tokens_seen": 4288280,
      "step": 6575
    },
    {
      "epoch": 3.448637316561845,
      "grad_norm": 0.3217565715312958,
      "learning_rate": 4.920575099679204e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 4291576,
      "step": 6580
    },
    {
      "epoch": 3.45125786163522,
      "grad_norm": 0.5374841094017029,
      "learning_rate": 4.920288919441383e-05,
      "loss": 0.4902,
      "num_input_tokens_seen": 4295288,
      "step": 6585
    },
    {
      "epoch": 3.4538784067085953,
      "grad_norm": 0.35167059302330017,
      "learning_rate": 4.9200022329067316e-05,
      "loss": 0.5409,
      "num_input_tokens_seen": 4298808,
      "step": 6590
    },
    {
      "epoch": 3.4564989517819704,
      "grad_norm": 0.6711329817771912,
      "learning_rate": 4.9197150401352224e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 4304280,
      "step": 6595
    },
    {
      "epoch": 3.459119496855346,
      "grad_norm": 0.7942778468132019,
      "learning_rate": 4.919427341186933e-05,
      "loss": 0.6933,
      "num_input_tokens_seen": 4306616,
      "step": 6600
    },
    {
      "epoch": 3.461740041928721,
      "grad_norm": 0.6627101898193359,
      "learning_rate": 4.919139136122045e-05,
      "loss": 0.4879,
      "num_input_tokens_seen": 4309112,
      "step": 6605
    },
    {
      "epoch": 3.4643605870020964,
      "grad_norm": 0.9737314581871033,
      "learning_rate": 4.918850425000849e-05,
      "loss": 0.5735,
      "num_input_tokens_seen": 4311832,
      "step": 6610
    },
    {
      "epoch": 3.4669811320754715,
      "grad_norm": 0.11246780306100845,
      "learning_rate": 4.918561207883741e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 4318328,
      "step": 6615
    },
    {
      "epoch": 3.469601677148847,
      "grad_norm": 0.6166829466819763,
      "learning_rate": 4.9182714848312195e-05,
      "loss": 0.4592,
      "num_input_tokens_seen": 4321848,
      "step": 6620
    },
    {
      "epoch": 3.4722222222222223,
      "grad_norm": 0.23418913781642914,
      "learning_rate": 4.917981255903893e-05,
      "loss": 0.3453,
      "num_input_tokens_seen": 4324984,
      "step": 6625
    },
    {
      "epoch": 3.4748427672955975,
      "grad_norm": 0.592635989189148,
      "learning_rate": 4.917690521162474e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 4327704,
      "step": 6630
    },
    {
      "epoch": 3.4774633123689727,
      "grad_norm": 0.34593793749809265,
      "learning_rate": 4.9173992806677796e-05,
      "loss": 0.4259,
      "num_input_tokens_seen": 4330904,
      "step": 6635
    },
    {
      "epoch": 3.480083857442348,
      "grad_norm": 0.7096487879753113,
      "learning_rate": 4.917107534480735e-05,
      "loss": 0.5355,
      "num_input_tokens_seen": 4333592,
      "step": 6640
    },
    {
      "epoch": 3.4827044025157234,
      "grad_norm": 0.6766663193702698,
      "learning_rate": 4.916815282662369e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 4336696,
      "step": 6645
    },
    {
      "epoch": 3.4853249475890986,
      "grad_norm": 0.9894431233406067,
      "learning_rate": 4.91652252527382e-05,
      "loss": 0.7407,
      "num_input_tokens_seen": 4340088,
      "step": 6650
    },
    {
      "epoch": 3.487945492662474,
      "grad_norm": 0.5787911415100098,
      "learning_rate": 4.9162292623763275e-05,
      "loss": 0.5494,
      "num_input_tokens_seen": 4343640,
      "step": 6655
    },
    {
      "epoch": 3.490566037735849,
      "grad_norm": 0.39999017119407654,
      "learning_rate": 4.9159354940312386e-05,
      "loss": 0.4601,
      "num_input_tokens_seen": 4348568,
      "step": 6660
    },
    {
      "epoch": 3.4931865828092246,
      "grad_norm": 0.6139646172523499,
      "learning_rate": 4.915641220300007e-05,
      "loss": 0.4239,
      "num_input_tokens_seen": 4352376,
      "step": 6665
    },
    {
      "epoch": 3.4958071278825997,
      "grad_norm": 0.3669530749320984,
      "learning_rate": 4.9153464412441915e-05,
      "loss": 0.5265,
      "num_input_tokens_seen": 4356120,
      "step": 6670
    },
    {
      "epoch": 3.498427672955975,
      "grad_norm": 0.5200740098953247,
      "learning_rate": 4.915051156925456e-05,
      "loss": 0.4908,
      "num_input_tokens_seen": 4359448,
      "step": 6675
    },
    {
      "epoch": 3.50104821802935,
      "grad_norm": 0.4257003664970398,
      "learning_rate": 4.91475536740557e-05,
      "loss": 0.3659,
      "num_input_tokens_seen": 4362936,
      "step": 6680
    },
    {
      "epoch": 3.5036687631027252,
      "grad_norm": 0.3794694244861603,
      "learning_rate": 4.914459072746411e-05,
      "loss": 0.4319,
      "num_input_tokens_seen": 4366968,
      "step": 6685
    },
    {
      "epoch": 3.5062893081761004,
      "grad_norm": 0.31786975264549255,
      "learning_rate": 4.91416227300996e-05,
      "loss": 0.373,
      "num_input_tokens_seen": 4370264,
      "step": 6690
    },
    {
      "epoch": 3.508909853249476,
      "grad_norm": 0.5786966681480408,
      "learning_rate": 4.913864968258303e-05,
      "loss": 0.3496,
      "num_input_tokens_seen": 4373432,
      "step": 6695
    },
    {
      "epoch": 3.511530398322851,
      "grad_norm": 0.4031195640563965,
      "learning_rate": 4.913567158553634e-05,
      "loss": 0.6256,
      "num_input_tokens_seen": 4376824,
      "step": 6700
    },
    {
      "epoch": 3.5141509433962264,
      "grad_norm": 0.3508337438106537,
      "learning_rate": 4.9132688439582505e-05,
      "loss": 0.5507,
      "num_input_tokens_seen": 4380504,
      "step": 6705
    },
    {
      "epoch": 3.5167714884696015,
      "grad_norm": 0.3681502044200897,
      "learning_rate": 4.9129700245345565e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 4383704,
      "step": 6710
    },
    {
      "epoch": 3.519392033542977,
      "grad_norm": 0.6192303895950317,
      "learning_rate": 4.912670700345063e-05,
      "loss": 0.6065,
      "num_input_tokens_seen": 4387128,
      "step": 6715
    },
    {
      "epoch": 3.5220125786163523,
      "grad_norm": 0.46538957953453064,
      "learning_rate": 4.912370871452383e-05,
      "loss": 0.5048,
      "num_input_tokens_seen": 4390296,
      "step": 6720
    },
    {
      "epoch": 3.5246331236897275,
      "grad_norm": 0.49749380350112915,
      "learning_rate": 4.91207053791924e-05,
      "loss": 0.5442,
      "num_input_tokens_seen": 4393624,
      "step": 6725
    },
    {
      "epoch": 3.5272536687631026,
      "grad_norm": 0.3797791004180908,
      "learning_rate": 4.911769699808457e-05,
      "loss": 0.4253,
      "num_input_tokens_seen": 4397944,
      "step": 6730
    },
    {
      "epoch": 3.529874213836478,
      "grad_norm": 0.36672496795654297,
      "learning_rate": 4.911468357182969e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 4400696,
      "step": 6735
    },
    {
      "epoch": 3.532494758909853,
      "grad_norm": 0.6325874328613281,
      "learning_rate": 4.9111665101058116e-05,
      "loss": 0.4835,
      "num_input_tokens_seen": 4403416,
      "step": 6740
    },
    {
      "epoch": 3.5351153039832286,
      "grad_norm": 0.33205440640449524,
      "learning_rate": 4.910864158640128e-05,
      "loss": 0.4225,
      "num_input_tokens_seen": 4407320,
      "step": 6745
    },
    {
      "epoch": 3.5377358490566038,
      "grad_norm": 0.5154252052307129,
      "learning_rate": 4.910561302849168e-05,
      "loss": 0.4516,
      "num_input_tokens_seen": 4409848,
      "step": 6750
    },
    {
      "epoch": 3.540356394129979,
      "grad_norm": 0.4606991112232208,
      "learning_rate": 4.910257942796284e-05,
      "loss": 0.4885,
      "num_input_tokens_seen": 4412920,
      "step": 6755
    },
    {
      "epoch": 3.5429769392033545,
      "grad_norm": 0.7083714008331299,
      "learning_rate": 4.909954078544936e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 4415384,
      "step": 6760
    },
    {
      "epoch": 3.5455974842767297,
      "grad_norm": 0.5376225113868713,
      "learning_rate": 4.909649710158689e-05,
      "loss": 0.3942,
      "num_input_tokens_seen": 4418040,
      "step": 6765
    },
    {
      "epoch": 3.548218029350105,
      "grad_norm": 0.32751697301864624,
      "learning_rate": 4.909344837701214e-05,
      "loss": 0.48,
      "num_input_tokens_seen": 4421464,
      "step": 6770
    },
    {
      "epoch": 3.55083857442348,
      "grad_norm": 0.4526107907295227,
      "learning_rate": 4.909039461236285e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 4423672,
      "step": 6775
    },
    {
      "epoch": 3.5534591194968552,
      "grad_norm": 0.47272154688835144,
      "learning_rate": 4.908733580827786e-05,
      "loss": 0.4411,
      "num_input_tokens_seen": 4426648,
      "step": 6780
    },
    {
      "epoch": 3.5560796645702304,
      "grad_norm": 0.637719988822937,
      "learning_rate": 4.9084271965397014e-05,
      "loss": 0.5168,
      "num_input_tokens_seen": 4429752,
      "step": 6785
    },
    {
      "epoch": 3.558700209643606,
      "grad_norm": 0.41394978761672974,
      "learning_rate": 4.908120308436124e-05,
      "loss": 0.6057,
      "num_input_tokens_seen": 4433784,
      "step": 6790
    },
    {
      "epoch": 3.561320754716981,
      "grad_norm": 0.6081231832504272,
      "learning_rate": 4.907812916581251e-05,
      "loss": 0.5028,
      "num_input_tokens_seen": 4436376,
      "step": 6795
    },
    {
      "epoch": 3.5639412997903563,
      "grad_norm": 0.630077064037323,
      "learning_rate": 4.907505021039386e-05,
      "loss": 0.4276,
      "num_input_tokens_seen": 4439384,
      "step": 6800
    },
    {
      "epoch": 3.5665618448637315,
      "grad_norm": 0.9931904077529907,
      "learning_rate": 4.907196621874937e-05,
      "loss": 0.5112,
      "num_input_tokens_seen": 4442712,
      "step": 6805
    },
    {
      "epoch": 3.569182389937107,
      "grad_norm": 0.6107714176177979,
      "learning_rate": 4.9068877191524176e-05,
      "loss": 0.519,
      "num_input_tokens_seen": 4446936,
      "step": 6810
    },
    {
      "epoch": 3.5718029350104823,
      "grad_norm": 0.5127824544906616,
      "learning_rate": 4.906578312936446e-05,
      "loss": 0.5443,
      "num_input_tokens_seen": 4450328,
      "step": 6815
    },
    {
      "epoch": 3.5744234800838575,
      "grad_norm": 0.40273770689964294,
      "learning_rate": 4.906268403291747e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 4453816,
      "step": 6820
    },
    {
      "epoch": 3.5770440251572326,
      "grad_norm": 0.34767013788223267,
      "learning_rate": 4.905957990283151e-05,
      "loss": 0.4055,
      "num_input_tokens_seen": 4458072,
      "step": 6825
    },
    {
      "epoch": 3.579664570230608,
      "grad_norm": 0.2694704830646515,
      "learning_rate": 4.905647073975592e-05,
      "loss": 0.4303,
      "num_input_tokens_seen": 4461144,
      "step": 6830
    },
    {
      "epoch": 3.582285115303983,
      "grad_norm": 0.38548412919044495,
      "learning_rate": 4.9053356544341095e-05,
      "loss": 0.3801,
      "num_input_tokens_seen": 4463928,
      "step": 6835
    },
    {
      "epoch": 3.5849056603773586,
      "grad_norm": 0.6823648810386658,
      "learning_rate": 4.90502373172385e-05,
      "loss": 0.4746,
      "num_input_tokens_seen": 4466488,
      "step": 6840
    },
    {
      "epoch": 3.5875262054507338,
      "grad_norm": 0.25229835510253906,
      "learning_rate": 4.9047113059100635e-05,
      "loss": 0.4149,
      "num_input_tokens_seen": 4470104,
      "step": 6845
    },
    {
      "epoch": 3.590146750524109,
      "grad_norm": 0.5833582282066345,
      "learning_rate": 4.904398377058106e-05,
      "loss": 0.5923,
      "num_input_tokens_seen": 4472184,
      "step": 6850
    },
    {
      "epoch": 3.5927672955974845,
      "grad_norm": 0.3650117814540863,
      "learning_rate": 4.90408494523344e-05,
      "loss": 0.4917,
      "num_input_tokens_seen": 4474840,
      "step": 6855
    },
    {
      "epoch": 3.5953878406708597,
      "grad_norm": 0.8952220678329468,
      "learning_rate": 4.90377101050163e-05,
      "loss": 0.562,
      "num_input_tokens_seen": 4477816,
      "step": 6860
    },
    {
      "epoch": 3.598008385744235,
      "grad_norm": 0.4216640591621399,
      "learning_rate": 4.903456572928348e-05,
      "loss": 0.56,
      "num_input_tokens_seen": 4481816,
      "step": 6865
    },
    {
      "epoch": 3.60062893081761,
      "grad_norm": 0.5364955067634583,
      "learning_rate": 4.903141632579371e-05,
      "loss": 0.3866,
      "num_input_tokens_seen": 4485464,
      "step": 6870
    },
    {
      "epoch": 3.603249475890985,
      "grad_norm": 0.3954927921295166,
      "learning_rate": 4.902826189520581e-05,
      "loss": 0.3925,
      "num_input_tokens_seen": 4490392,
      "step": 6875
    },
    {
      "epoch": 3.6058700209643604,
      "grad_norm": 0.5638694167137146,
      "learning_rate": 4.9025102438179656e-05,
      "loss": 0.5047,
      "num_input_tokens_seen": 4494712,
      "step": 6880
    },
    {
      "epoch": 3.608490566037736,
      "grad_norm": 0.38750559091567993,
      "learning_rate": 4.902193795537616e-05,
      "loss": 0.5393,
      "num_input_tokens_seen": 4497528,
      "step": 6885
    },
    {
      "epoch": 3.611111111111111,
      "grad_norm": 0.3684803545475006,
      "learning_rate": 4.90187684474573e-05,
      "loss": 0.4006,
      "num_input_tokens_seen": 4500536,
      "step": 6890
    },
    {
      "epoch": 3.6137316561844863,
      "grad_norm": 0.4240759611129761,
      "learning_rate": 4.901559391508611e-05,
      "loss": 0.5164,
      "num_input_tokens_seen": 4504056,
      "step": 6895
    },
    {
      "epoch": 3.6163522012578615,
      "grad_norm": 0.40147513151168823,
      "learning_rate": 4.9012414358926644e-05,
      "loss": 0.4643,
      "num_input_tokens_seen": 4506936,
      "step": 6900
    },
    {
      "epoch": 3.618972746331237,
      "grad_norm": 0.4136345684528351,
      "learning_rate": 4.900922977964405e-05,
      "loss": 0.3818,
      "num_input_tokens_seen": 4509560,
      "step": 6905
    },
    {
      "epoch": 3.6215932914046123,
      "grad_norm": 0.5175248384475708,
      "learning_rate": 4.90060401779045e-05,
      "loss": 0.482,
      "num_input_tokens_seen": 4513240,
      "step": 6910
    },
    {
      "epoch": 3.6242138364779874,
      "grad_norm": 0.4691300690174103,
      "learning_rate": 4.900284555437521e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 4515640,
      "step": 6915
    },
    {
      "epoch": 3.6268343815513626,
      "grad_norm": 0.2926447093486786,
      "learning_rate": 4.899964590972448e-05,
      "loss": 0.4158,
      "num_input_tokens_seen": 4520824,
      "step": 6920
    },
    {
      "epoch": 3.629454926624738,
      "grad_norm": 0.4195452630519867,
      "learning_rate": 4.899644124462163e-05,
      "loss": 0.5059,
      "num_input_tokens_seen": 4523896,
      "step": 6925
    },
    {
      "epoch": 3.632075471698113,
      "grad_norm": 0.5352759957313538,
      "learning_rate": 4.8993231559737026e-05,
      "loss": 0.6224,
      "num_input_tokens_seen": 4527000,
      "step": 6930
    },
    {
      "epoch": 3.6346960167714886,
      "grad_norm": 0.6202412843704224,
      "learning_rate": 4.8990016855742114e-05,
      "loss": 0.5987,
      "num_input_tokens_seen": 4529912,
      "step": 6935
    },
    {
      "epoch": 3.6373165618448637,
      "grad_norm": 0.6824647784233093,
      "learning_rate": 4.898679713330937e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 4533336,
      "step": 6940
    },
    {
      "epoch": 3.639937106918239,
      "grad_norm": 0.38813379406929016,
      "learning_rate": 4.898357239311232e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 4536344,
      "step": 6945
    },
    {
      "epoch": 3.6425576519916145,
      "grad_norm": 0.3057459890842438,
      "learning_rate": 4.898034263582554e-05,
      "loss": 0.4243,
      "num_input_tokens_seen": 4540152,
      "step": 6950
    },
    {
      "epoch": 3.6451781970649897,
      "grad_norm": 0.4968242347240448,
      "learning_rate": 4.897710786212467e-05,
      "loss": 0.359,
      "num_input_tokens_seen": 4543032,
      "step": 6955
    },
    {
      "epoch": 3.647798742138365,
      "grad_norm": 0.708433210849762,
      "learning_rate": 4.897386807268638e-05,
      "loss": 0.4864,
      "num_input_tokens_seen": 4547064,
      "step": 6960
    },
    {
      "epoch": 3.65041928721174,
      "grad_norm": 0.40982240438461304,
      "learning_rate": 4.897062326818839e-05,
      "loss": 0.5261,
      "num_input_tokens_seen": 4551000,
      "step": 6965
    },
    {
      "epoch": 3.653039832285115,
      "grad_norm": 0.28248605132102966,
      "learning_rate": 4.8967373449309495e-05,
      "loss": 0.5469,
      "num_input_tokens_seen": 4555000,
      "step": 6970
    },
    {
      "epoch": 3.6556603773584904,
      "grad_norm": 0.22936265170574188,
      "learning_rate": 4.89641186167295e-05,
      "loss": 0.432,
      "num_input_tokens_seen": 4558680,
      "step": 6975
    },
    {
      "epoch": 3.658280922431866,
      "grad_norm": 0.4240109324455261,
      "learning_rate": 4.8960858771129294e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 4561336,
      "step": 6980
    },
    {
      "epoch": 3.660901467505241,
      "grad_norm": 0.34136202931404114,
      "learning_rate": 4.895759391319079e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 4564568,
      "step": 6985
    },
    {
      "epoch": 3.6635220125786163,
      "grad_norm": 0.487051784992218,
      "learning_rate": 4.8954324043596954e-05,
      "loss": 0.4544,
      "num_input_tokens_seen": 4567448,
      "step": 6990
    },
    {
      "epoch": 3.6661425576519915,
      "grad_norm": 0.2840437889099121,
      "learning_rate": 4.8951049163031826e-05,
      "loss": 0.5821,
      "num_input_tokens_seen": 4570136,
      "step": 6995
    },
    {
      "epoch": 3.668763102725367,
      "grad_norm": 0.5121292471885681,
      "learning_rate": 4.894776927218045e-05,
      "loss": 0.4927,
      "num_input_tokens_seen": 4573624,
      "step": 7000
    },
    {
      "epoch": 3.6713836477987423,
      "grad_norm": 1.3381385803222656,
      "learning_rate": 4.894448437172896e-05,
      "loss": 0.4352,
      "num_input_tokens_seen": 4577048,
      "step": 7005
    },
    {
      "epoch": 3.6740041928721174,
      "grad_norm": 0.3929733335971832,
      "learning_rate": 4.894119446236452e-05,
      "loss": 0.5254,
      "num_input_tokens_seen": 4580088,
      "step": 7010
    },
    {
      "epoch": 3.6766247379454926,
      "grad_norm": 0.3874812424182892,
      "learning_rate": 4.893789954477532e-05,
      "loss": 0.4127,
      "num_input_tokens_seen": 4583480,
      "step": 7015
    },
    {
      "epoch": 3.6792452830188678,
      "grad_norm": 0.5091656446456909,
      "learning_rate": 4.893459961965063e-05,
      "loss": 0.6544,
      "num_input_tokens_seen": 4586168,
      "step": 7020
    },
    {
      "epoch": 3.681865828092243,
      "grad_norm": 0.35908475518226624,
      "learning_rate": 4.8931294687680775e-05,
      "loss": 0.3964,
      "num_input_tokens_seen": 4589592,
      "step": 7025
    },
    {
      "epoch": 3.6844863731656186,
      "grad_norm": 0.7898173332214355,
      "learning_rate": 4.8927984749557086e-05,
      "loss": 0.5624,
      "num_input_tokens_seen": 4592440,
      "step": 7030
    },
    {
      "epoch": 3.6871069182389937,
      "grad_norm": 0.4831688404083252,
      "learning_rate": 4.892466980597198e-05,
      "loss": 0.3656,
      "num_input_tokens_seen": 4594904,
      "step": 7035
    },
    {
      "epoch": 3.689727463312369,
      "grad_norm": 0.4256344437599182,
      "learning_rate": 4.892134985761889e-05,
      "loss": 0.542,
      "num_input_tokens_seen": 4597624,
      "step": 7040
    },
    {
      "epoch": 3.6923480083857445,
      "grad_norm": 0.7184138298034668,
      "learning_rate": 4.891802490519232e-05,
      "loss": 0.5632,
      "num_input_tokens_seen": 4599928,
      "step": 7045
    },
    {
      "epoch": 3.6949685534591197,
      "grad_norm": 1.0586222410202026,
      "learning_rate": 4.891469494938781e-05,
      "loss": 0.4959,
      "num_input_tokens_seen": 4602328,
      "step": 7050
    },
    {
      "epoch": 3.697589098532495,
      "grad_norm": 0.4512532651424408,
      "learning_rate": 4.891135999090195e-05,
      "loss": 0.5076,
      "num_input_tokens_seen": 4606328,
      "step": 7055
    },
    {
      "epoch": 3.70020964360587,
      "grad_norm": 0.3982628881931305,
      "learning_rate": 4.890802003043238e-05,
      "loss": 0.3924,
      "num_input_tokens_seen": 4609176,
      "step": 7060
    },
    {
      "epoch": 3.702830188679245,
      "grad_norm": 0.3468991219997406,
      "learning_rate": 4.890467506867777e-05,
      "loss": 0.5651,
      "num_input_tokens_seen": 4612056,
      "step": 7065
    },
    {
      "epoch": 3.7054507337526204,
      "grad_norm": 0.581133246421814,
      "learning_rate": 4.890132510633786e-05,
      "loss": 0.6071,
      "num_input_tokens_seen": 4615320,
      "step": 7070
    },
    {
      "epoch": 3.708071278825996,
      "grad_norm": 0.4326293468475342,
      "learning_rate": 4.8897970144113416e-05,
      "loss": 0.5583,
      "num_input_tokens_seen": 4618200,
      "step": 7075
    },
    {
      "epoch": 3.710691823899371,
      "grad_norm": 0.6214519739151001,
      "learning_rate": 4.889461018270626e-05,
      "loss": 0.354,
      "num_input_tokens_seen": 4621368,
      "step": 7080
    },
    {
      "epoch": 3.7133123689727463,
      "grad_norm": 0.4311123490333557,
      "learning_rate": 4.889124522281926e-05,
      "loss": 0.5102,
      "num_input_tokens_seen": 4624088,
      "step": 7085
    },
    {
      "epoch": 3.7159329140461215,
      "grad_norm": 0.3731737732887268,
      "learning_rate": 4.888787526515632e-05,
      "loss": 0.6205,
      "num_input_tokens_seen": 4628568,
      "step": 7090
    },
    {
      "epoch": 3.718553459119497,
      "grad_norm": 0.4406973123550415,
      "learning_rate": 4.8884500310422406e-05,
      "loss": 0.5333,
      "num_input_tokens_seen": 4631288,
      "step": 7095
    },
    {
      "epoch": 3.7211740041928723,
      "grad_norm": 0.2925749123096466,
      "learning_rate": 4.8881120359323516e-05,
      "loss": 0.3954,
      "num_input_tokens_seen": 4635224,
      "step": 7100
    },
    {
      "epoch": 3.7237945492662474,
      "grad_norm": 0.4186607003211975,
      "learning_rate": 4.887773541256669e-05,
      "loss": 0.705,
      "num_input_tokens_seen": 4638648,
      "step": 7105
    },
    {
      "epoch": 3.7264150943396226,
      "grad_norm": 0.747218906879425,
      "learning_rate": 4.887434547086004e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 4641848,
      "step": 7110
    },
    {
      "epoch": 3.7290356394129978,
      "grad_norm": 0.4153288006782532,
      "learning_rate": 4.887095053491269e-05,
      "loss": 0.5029,
      "num_input_tokens_seen": 4644408,
      "step": 7115
    },
    {
      "epoch": 3.731656184486373,
      "grad_norm": 0.5113476514816284,
      "learning_rate": 4.8867550605434806e-05,
      "loss": 0.6337,
      "num_input_tokens_seen": 4647096,
      "step": 7120
    },
    {
      "epoch": 3.7342767295597485,
      "grad_norm": 0.42867180705070496,
      "learning_rate": 4.8864145683137644e-05,
      "loss": 0.4954,
      "num_input_tokens_seen": 4649176,
      "step": 7125
    },
    {
      "epoch": 3.7368972746331237,
      "grad_norm": 0.4620937705039978,
      "learning_rate": 4.886073576873346e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 4653144,
      "step": 7130
    },
    {
      "epoch": 3.739517819706499,
      "grad_norm": 0.3885064423084259,
      "learning_rate": 4.885732086293556e-05,
      "loss": 0.3627,
      "num_input_tokens_seen": 4655864,
      "step": 7135
    },
    {
      "epoch": 3.742138364779874,
      "grad_norm": 0.385978639125824,
      "learning_rate": 4.8853900966458335e-05,
      "loss": 0.5001,
      "num_input_tokens_seen": 4659000,
      "step": 7140
    },
    {
      "epoch": 3.7447589098532497,
      "grad_norm": 0.283728688955307,
      "learning_rate": 4.8850476080017146e-05,
      "loss": 0.5387,
      "num_input_tokens_seen": 4661720,
      "step": 7145
    },
    {
      "epoch": 3.747379454926625,
      "grad_norm": 0.8210864663124084,
      "learning_rate": 4.884704620432848e-05,
      "loss": 0.4687,
      "num_input_tokens_seen": 4666584,
      "step": 7150
    },
    {
      "epoch": 3.75,
      "grad_norm": 0.5698733329772949,
      "learning_rate": 4.8843611340109794e-05,
      "loss": 0.4705,
      "num_input_tokens_seen": 4669272,
      "step": 7155
    },
    {
      "epoch": 3.752620545073375,
      "grad_norm": 0.6524546146392822,
      "learning_rate": 4.884017148807965e-05,
      "loss": 0.4943,
      "num_input_tokens_seen": 4672280,
      "step": 7160
    },
    {
      "epoch": 3.7552410901467503,
      "grad_norm": 0.47294795513153076,
      "learning_rate": 4.883672664895761e-05,
      "loss": 0.3692,
      "num_input_tokens_seen": 4675032,
      "step": 7165
    },
    {
      "epoch": 3.757861635220126,
      "grad_norm": 0.4427131414413452,
      "learning_rate": 4.8833276823464296e-05,
      "loss": 0.4701,
      "num_input_tokens_seen": 4678360,
      "step": 7170
    },
    {
      "epoch": 3.760482180293501,
      "grad_norm": 1.302556037902832,
      "learning_rate": 4.882982201232138e-05,
      "loss": 0.5094,
      "num_input_tokens_seen": 4681208,
      "step": 7175
    },
    {
      "epoch": 3.7631027253668763,
      "grad_norm": 0.3091582953929901,
      "learning_rate": 4.882636221625156e-05,
      "loss": 0.4766,
      "num_input_tokens_seen": 4683768,
      "step": 7180
    },
    {
      "epoch": 3.7657232704402515,
      "grad_norm": 0.7563947439193726,
      "learning_rate": 4.8822897435978586e-05,
      "loss": 0.4894,
      "num_input_tokens_seen": 4686872,
      "step": 7185
    },
    {
      "epoch": 3.768343815513627,
      "grad_norm": 0.3171437084674835,
      "learning_rate": 4.8819427672227264e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 4690424,
      "step": 7190
    },
    {
      "epoch": 3.7709643605870022,
      "grad_norm": 0.6367666721343994,
      "learning_rate": 4.881595292572341e-05,
      "loss": 0.5595,
      "num_input_tokens_seen": 4692824,
      "step": 7195
    },
    {
      "epoch": 3.7735849056603774,
      "grad_norm": 0.4441646933555603,
      "learning_rate": 4.881247319719392e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 4696440,
      "step": 7200
    },
    {
      "epoch": 3.7762054507337526,
      "grad_norm": 0.46833375096321106,
      "learning_rate": 4.88089884873667e-05,
      "loss": 0.4337,
      "num_input_tokens_seen": 4700440,
      "step": 7205
    },
    {
      "epoch": 3.7788259958071277,
      "grad_norm": 0.36314719915390015,
      "learning_rate": 4.880549879697071e-05,
      "loss": 0.4528,
      "num_input_tokens_seen": 4703320,
      "step": 7210
    },
    {
      "epoch": 3.781446540880503,
      "grad_norm": 0.40450283885002136,
      "learning_rate": 4.8802004126735965e-05,
      "loss": 0.5745,
      "num_input_tokens_seen": 4706552,
      "step": 7215
    },
    {
      "epoch": 3.7840670859538785,
      "grad_norm": 0.23910818994045258,
      "learning_rate": 4.87985044773935e-05,
      "loss": 0.4112,
      "num_input_tokens_seen": 4709592,
      "step": 7220
    },
    {
      "epoch": 3.7866876310272537,
      "grad_norm": 0.5090042948722839,
      "learning_rate": 4.879499984967541e-05,
      "loss": 0.5817,
      "num_input_tokens_seen": 4712344,
      "step": 7225
    },
    {
      "epoch": 3.789308176100629,
      "grad_norm": 0.6871979832649231,
      "learning_rate": 4.8791490244314824e-05,
      "loss": 0.5453,
      "num_input_tokens_seen": 4715000,
      "step": 7230
    },
    {
      "epoch": 3.791928721174004,
      "grad_norm": 0.46084025502204895,
      "learning_rate": 4.8787975662045904e-05,
      "loss": 0.4463,
      "num_input_tokens_seen": 4718904,
      "step": 7235
    },
    {
      "epoch": 3.7945492662473796,
      "grad_norm": 0.62574702501297,
      "learning_rate": 4.878445610360386e-05,
      "loss": 0.5801,
      "num_input_tokens_seen": 4721656,
      "step": 7240
    },
    {
      "epoch": 3.797169811320755,
      "grad_norm": 1.6257692575454712,
      "learning_rate": 4.878093156972494e-05,
      "loss": 0.5733,
      "num_input_tokens_seen": 4726264,
      "step": 7245
    },
    {
      "epoch": 3.79979035639413,
      "grad_norm": 0.4668174088001251,
      "learning_rate": 4.877740206114645e-05,
      "loss": 0.43,
      "num_input_tokens_seen": 4729112,
      "step": 7250
    },
    {
      "epoch": 3.802410901467505,
      "grad_norm": 0.4225127100944519,
      "learning_rate": 4.8773867578606716e-05,
      "loss": 0.5473,
      "num_input_tokens_seen": 4731832,
      "step": 7255
    },
    {
      "epoch": 3.8050314465408803,
      "grad_norm": 0.4891183078289032,
      "learning_rate": 4.877032812284511e-05,
      "loss": 0.5912,
      "num_input_tokens_seen": 4738136,
      "step": 7260
    },
    {
      "epoch": 3.8076519916142555,
      "grad_norm": 0.3685946464538574,
      "learning_rate": 4.876678369460205e-05,
      "loss": 0.425,
      "num_input_tokens_seen": 4741080,
      "step": 7265
    },
    {
      "epoch": 3.810272536687631,
      "grad_norm": 0.3917217254638672,
      "learning_rate": 4.8763234294618985e-05,
      "loss": 0.3886,
      "num_input_tokens_seen": 4744472,
      "step": 7270
    },
    {
      "epoch": 3.8128930817610063,
      "grad_norm": 0.3712019622325897,
      "learning_rate": 4.875967992363841e-05,
      "loss": 0.394,
      "num_input_tokens_seen": 4748664,
      "step": 7275
    },
    {
      "epoch": 3.8155136268343814,
      "grad_norm": 0.7707312107086182,
      "learning_rate": 4.875612058240386e-05,
      "loss": 0.5045,
      "num_input_tokens_seen": 4751704,
      "step": 7280
    },
    {
      "epoch": 3.818134171907757,
      "grad_norm": 2.0317769050598145,
      "learning_rate": 4.875255627165991e-05,
      "loss": 0.5566,
      "num_input_tokens_seen": 4754104,
      "step": 7285
    },
    {
      "epoch": 3.8207547169811322,
      "grad_norm": 0.4176481366157532,
      "learning_rate": 4.874898699215217e-05,
      "loss": 0.4425,
      "num_input_tokens_seen": 4757528,
      "step": 7290
    },
    {
      "epoch": 3.8233752620545074,
      "grad_norm": 0.28168970346450806,
      "learning_rate": 4.8745412744627296e-05,
      "loss": 0.6254,
      "num_input_tokens_seen": 4761208,
      "step": 7295
    },
    {
      "epoch": 3.8259958071278826,
      "grad_norm": 0.4516281485557556,
      "learning_rate": 4.874183352983297e-05,
      "loss": 0.4666,
      "num_input_tokens_seen": 4764088,
      "step": 7300
    },
    {
      "epoch": 3.8286163522012577,
      "grad_norm": 0.3010540306568146,
      "learning_rate": 4.873824934851794e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 4767352,
      "step": 7305
    },
    {
      "epoch": 3.831236897274633,
      "grad_norm": 0.7222611904144287,
      "learning_rate": 4.873466020143196e-05,
      "loss": 0.4912,
      "num_input_tokens_seen": 4770648,
      "step": 7310
    },
    {
      "epoch": 3.8338574423480085,
      "grad_norm": 0.32598063349723816,
      "learning_rate": 4.873106608932585e-05,
      "loss": 0.4789,
      "num_input_tokens_seen": 4777624,
      "step": 7315
    },
    {
      "epoch": 3.8364779874213837,
      "grad_norm": 0.6053031086921692,
      "learning_rate": 4.872746701295145e-05,
      "loss": 0.5657,
      "num_input_tokens_seen": 4780376,
      "step": 7320
    },
    {
      "epoch": 3.839098532494759,
      "grad_norm": 0.2988950312137604,
      "learning_rate": 4.872386297306165e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 4783096,
      "step": 7325
    },
    {
      "epoch": 3.841719077568134,
      "grad_norm": 0.3212558627128601,
      "learning_rate": 4.872025397041037e-05,
      "loss": 0.3465,
      "num_input_tokens_seen": 4787000,
      "step": 7330
    },
    {
      "epoch": 3.8443396226415096,
      "grad_norm": 0.828147292137146,
      "learning_rate": 4.871664000575258e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 4790168,
      "step": 7335
    },
    {
      "epoch": 3.846960167714885,
      "grad_norm": 0.3340201675891876,
      "learning_rate": 4.871302107984427e-05,
      "loss": 0.5393,
      "num_input_tokens_seen": 4793944,
      "step": 7340
    },
    {
      "epoch": 3.84958071278826,
      "grad_norm": 0.453649640083313,
      "learning_rate": 4.8709397193442474e-05,
      "loss": 0.4354,
      "num_input_tokens_seen": 4797016,
      "step": 7345
    },
    {
      "epoch": 3.852201257861635,
      "grad_norm": 0.27841517329216003,
      "learning_rate": 4.870576834730529e-05,
      "loss": 0.4816,
      "num_input_tokens_seen": 4800216,
      "step": 7350
    },
    {
      "epoch": 3.8548218029350103,
      "grad_norm": 0.3930520713329315,
      "learning_rate": 4.8702134542191816e-05,
      "loss": 0.3951,
      "num_input_tokens_seen": 4803992,
      "step": 7355
    },
    {
      "epoch": 3.8574423480083855,
      "grad_norm": 0.6236428022384644,
      "learning_rate": 4.86984957788622e-05,
      "loss": 0.4652,
      "num_input_tokens_seen": 4806872,
      "step": 7360
    },
    {
      "epoch": 3.860062893081761,
      "grad_norm": 0.43887850642204285,
      "learning_rate": 4.869485205807763e-05,
      "loss": 0.3039,
      "num_input_tokens_seen": 4809912,
      "step": 7365
    },
    {
      "epoch": 3.8626834381551363,
      "grad_norm": 0.30869537591934204,
      "learning_rate": 4.869120338060034e-05,
      "loss": 0.3399,
      "num_input_tokens_seen": 4816984,
      "step": 7370
    },
    {
      "epoch": 3.8653039832285114,
      "grad_norm": 0.35403376817703247,
      "learning_rate": 4.8687549747193585e-05,
      "loss": 0.5444,
      "num_input_tokens_seen": 4819608,
      "step": 7375
    },
    {
      "epoch": 3.867924528301887,
      "grad_norm": 0.20158714056015015,
      "learning_rate": 4.868389115862166e-05,
      "loss": 0.3991,
      "num_input_tokens_seen": 4823448,
      "step": 7380
    },
    {
      "epoch": 3.870545073375262,
      "grad_norm": 0.5213207602500916,
      "learning_rate": 4.868022761564991e-05,
      "loss": 0.5106,
      "num_input_tokens_seen": 4826328,
      "step": 7385
    },
    {
      "epoch": 3.8731656184486374,
      "grad_norm": 0.559551477432251,
      "learning_rate": 4.8676559119044696e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 4829048,
      "step": 7390
    },
    {
      "epoch": 3.8757861635220126,
      "grad_norm": 0.42469069361686707,
      "learning_rate": 4.867288566957344e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 4833048,
      "step": 7395
    },
    {
      "epoch": 3.8784067085953877,
      "grad_norm": 0.2970190644264221,
      "learning_rate": 4.866920726800457e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 4836184,
      "step": 7400
    },
    {
      "epoch": 3.881027253668763,
      "grad_norm": 0.8261346817016602,
      "learning_rate": 4.8665523915107576e-05,
      "loss": 0.4919,
      "num_input_tokens_seen": 4839896,
      "step": 7405
    },
    {
      "epoch": 3.8836477987421385,
      "grad_norm": 0.9560166001319885,
      "learning_rate": 4.866183561165296e-05,
      "loss": 0.4382,
      "num_input_tokens_seen": 4842936,
      "step": 7410
    },
    {
      "epoch": 3.8862683438155137,
      "grad_norm": 0.5760504603385925,
      "learning_rate": 4.86581423584123e-05,
      "loss": 0.6353,
      "num_input_tokens_seen": 4846200,
      "step": 7415
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 0.41748377680778503,
      "learning_rate": 4.8654444156158154e-05,
      "loss": 0.4795,
      "num_input_tokens_seen": 4849336,
      "step": 7420
    },
    {
      "epoch": 3.891509433962264,
      "grad_norm": 0.5455383658409119,
      "learning_rate": 4.865074100566415e-05,
      "loss": 0.4919,
      "num_input_tokens_seen": 4852632,
      "step": 7425
    },
    {
      "epoch": 3.8941299790356396,
      "grad_norm": 0.5664611458778381,
      "learning_rate": 4.864703290770495e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 4855384,
      "step": 7430
    },
    {
      "epoch": 3.896750524109015,
      "grad_norm": 0.46650445461273193,
      "learning_rate": 4.864331986305626e-05,
      "loss": 0.5185,
      "num_input_tokens_seen": 4858936,
      "step": 7435
    },
    {
      "epoch": 3.89937106918239,
      "grad_norm": 0.552236020565033,
      "learning_rate": 4.8639601872494775e-05,
      "loss": 0.5075,
      "num_input_tokens_seen": 4862488,
      "step": 7440
    },
    {
      "epoch": 3.901991614255765,
      "grad_norm": 0.38998252153396606,
      "learning_rate": 4.8635878936798287e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 4865240,
      "step": 7445
    },
    {
      "epoch": 3.9046121593291403,
      "grad_norm": 0.3191874027252197,
      "learning_rate": 4.863215105674557e-05,
      "loss": 0.461,
      "num_input_tokens_seen": 4868152,
      "step": 7450
    },
    {
      "epoch": 3.9072327044025155,
      "grad_norm": 0.5270214080810547,
      "learning_rate": 4.8628418233116457e-05,
      "loss": 0.3699,
      "num_input_tokens_seen": 4870808,
      "step": 7455
    },
    {
      "epoch": 3.909853249475891,
      "grad_norm": 0.3070122003555298,
      "learning_rate": 4.8624680466691824e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 4873944,
      "step": 7460
    },
    {
      "epoch": 3.9124737945492662,
      "grad_norm": 0.5357851386070251,
      "learning_rate": 4.862093775825356e-05,
      "loss": 0.5514,
      "num_input_tokens_seen": 4877528,
      "step": 7465
    },
    {
      "epoch": 3.9150943396226414,
      "grad_norm": 0.4665209650993347,
      "learning_rate": 4.86171901085846e-05,
      "loss": 0.4356,
      "num_input_tokens_seen": 4880504,
      "step": 7470
    },
    {
      "epoch": 3.917714884696017,
      "grad_norm": 0.48029664158821106,
      "learning_rate": 4.861343751846891e-05,
      "loss": 0.469,
      "num_input_tokens_seen": 4884312,
      "step": 7475
    },
    {
      "epoch": 3.920335429769392,
      "grad_norm": 0.5659381151199341,
      "learning_rate": 4.860967998869148e-05,
      "loss": 0.3819,
      "num_input_tokens_seen": 4887032,
      "step": 7480
    },
    {
      "epoch": 3.9229559748427674,
      "grad_norm": 0.4443244934082031,
      "learning_rate": 4.860591752003836e-05,
      "loss": 0.5272,
      "num_input_tokens_seen": 4889976,
      "step": 7485
    },
    {
      "epoch": 3.9255765199161425,
      "grad_norm": 0.2807394564151764,
      "learning_rate": 4.8602150113296596e-05,
      "loss": 0.5626,
      "num_input_tokens_seen": 4892600,
      "step": 7490
    },
    {
      "epoch": 3.9281970649895177,
      "grad_norm": 0.8591745495796204,
      "learning_rate": 4.8598377769254305e-05,
      "loss": 0.4664,
      "num_input_tokens_seen": 4895640,
      "step": 7495
    },
    {
      "epoch": 3.930817610062893,
      "grad_norm": 0.40002691745758057,
      "learning_rate": 4.85946004887006e-05,
      "loss": 0.4006,
      "num_input_tokens_seen": 4898200,
      "step": 7500
    },
    {
      "epoch": 3.9334381551362685,
      "grad_norm": 0.747114896774292,
      "learning_rate": 4.859081827242566e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 4901080,
      "step": 7505
    },
    {
      "epoch": 3.9360587002096437,
      "grad_norm": 0.38686296343803406,
      "learning_rate": 4.858703112122068e-05,
      "loss": 0.5218,
      "num_input_tokens_seen": 4904280,
      "step": 7510
    },
    {
      "epoch": 3.938679245283019,
      "grad_norm": 0.4964176118373871,
      "learning_rate": 4.8583239035877876e-05,
      "loss": 0.7063,
      "num_input_tokens_seen": 4907832,
      "step": 7515
    },
    {
      "epoch": 3.941299790356394,
      "grad_norm": 0.6622804403305054,
      "learning_rate": 4.8579442017190526e-05,
      "loss": 0.6298,
      "num_input_tokens_seen": 4911320,
      "step": 7520
    },
    {
      "epoch": 3.9439203354297696,
      "grad_norm": 0.262911319732666,
      "learning_rate": 4.8575640065952905e-05,
      "loss": 0.4656,
      "num_input_tokens_seen": 4914072,
      "step": 7525
    },
    {
      "epoch": 3.9465408805031448,
      "grad_norm": 0.5232764482498169,
      "learning_rate": 4.857183318296035e-05,
      "loss": 0.5129,
      "num_input_tokens_seen": 4918584,
      "step": 7530
    },
    {
      "epoch": 3.94916142557652,
      "grad_norm": 0.4943792521953583,
      "learning_rate": 4.856802136900922e-05,
      "loss": 0.4721,
      "num_input_tokens_seen": 4921496,
      "step": 7535
    },
    {
      "epoch": 3.951781970649895,
      "grad_norm": 0.5081577897071838,
      "learning_rate": 4.856420462489689e-05,
      "loss": 0.513,
      "num_input_tokens_seen": 4924600,
      "step": 7540
    },
    {
      "epoch": 3.9544025157232703,
      "grad_norm": 0.3118322491645813,
      "learning_rate": 4.856038295142179e-05,
      "loss": 0.4654,
      "num_input_tokens_seen": 4927512,
      "step": 7545
    },
    {
      "epoch": 3.9570230607966455,
      "grad_norm": 0.4935179352760315,
      "learning_rate": 4.855655634938337e-05,
      "loss": 0.4529,
      "num_input_tokens_seen": 4930360,
      "step": 7550
    },
    {
      "epoch": 3.959643605870021,
      "grad_norm": 0.3662487864494324,
      "learning_rate": 4.8552724819582106e-05,
      "loss": 0.3915,
      "num_input_tokens_seen": 4933496,
      "step": 7555
    },
    {
      "epoch": 3.9622641509433962,
      "grad_norm": 0.38100650906562805,
      "learning_rate": 4.8548888362819515e-05,
      "loss": 0.3959,
      "num_input_tokens_seen": 4936152,
      "step": 7560
    },
    {
      "epoch": 3.9648846960167714,
      "grad_norm": 0.7377107739448547,
      "learning_rate": 4.854504697989814e-05,
      "loss": 0.6908,
      "num_input_tokens_seen": 4939832,
      "step": 7565
    },
    {
      "epoch": 3.967505241090147,
      "grad_norm": 0.3941519856452942,
      "learning_rate": 4.8541200671621554e-05,
      "loss": 0.5227,
      "num_input_tokens_seen": 4943192,
      "step": 7570
    },
    {
      "epoch": 3.970125786163522,
      "grad_norm": 0.6633902192115784,
      "learning_rate": 4.8537349438794354e-05,
      "loss": 0.4567,
      "num_input_tokens_seen": 4946392,
      "step": 7575
    },
    {
      "epoch": 3.9727463312368974,
      "grad_norm": 0.32599809765815735,
      "learning_rate": 4.853349328222219e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 4949464,
      "step": 7580
    },
    {
      "epoch": 3.9753668763102725,
      "grad_norm": 0.4386586546897888,
      "learning_rate": 4.8529632202711706e-05,
      "loss": 0.3963,
      "num_input_tokens_seen": 4955288,
      "step": 7585
    },
    {
      "epoch": 3.9779874213836477,
      "grad_norm": 0.449459969997406,
      "learning_rate": 4.852576620107061e-05,
      "loss": 0.395,
      "num_input_tokens_seen": 4958008,
      "step": 7590
    },
    {
      "epoch": 3.980607966457023,
      "grad_norm": 0.3005140423774719,
      "learning_rate": 4.8521895278107615e-05,
      "loss": 0.3643,
      "num_input_tokens_seen": 4961400,
      "step": 7595
    },
    {
      "epoch": 3.9832285115303985,
      "grad_norm": 0.6274169683456421,
      "learning_rate": 4.8518019434632484e-05,
      "loss": 0.585,
      "num_input_tokens_seen": 4964536,
      "step": 7600
    },
    {
      "epoch": 3.9858490566037736,
      "grad_norm": 0.4730893075466156,
      "learning_rate": 4.8514138671455995e-05,
      "loss": 0.598,
      "num_input_tokens_seen": 4968280,
      "step": 7605
    },
    {
      "epoch": 3.988469601677149,
      "grad_norm": 0.32692426443099976,
      "learning_rate": 4.851025298938996e-05,
      "loss": 0.5732,
      "num_input_tokens_seen": 4971640,
      "step": 7610
    },
    {
      "epoch": 3.991090146750524,
      "grad_norm": 0.3517538607120514,
      "learning_rate": 4.850636238924722e-05,
      "loss": 0.5014,
      "num_input_tokens_seen": 4975512,
      "step": 7615
    },
    {
      "epoch": 3.9937106918238996,
      "grad_norm": 0.4339066743850708,
      "learning_rate": 4.850246687184163e-05,
      "loss": 0.5772,
      "num_input_tokens_seen": 4978520,
      "step": 7620
    },
    {
      "epoch": 3.9963312368972748,
      "grad_norm": 0.46479347348213196,
      "learning_rate": 4.849856643798812e-05,
      "loss": 0.4089,
      "num_input_tokens_seen": 4981944,
      "step": 7625
    },
    {
      "epoch": 3.99895178197065,
      "grad_norm": 0.6368457078933716,
      "learning_rate": 4.849466108850259e-05,
      "loss": 0.4343,
      "num_input_tokens_seen": 4984632,
      "step": 7630
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.49366217851638794,
      "eval_runtime": 9.7754,
      "eval_samples_per_second": 86.748,
      "eval_steps_per_second": 21.687,
      "num_input_tokens_seen": 4985200,
      "step": 7632
    },
    {
      "epoch": 4.001572327044025,
      "grad_norm": 0.3856513202190399,
      "learning_rate": 4.8490750824202e-05,
      "loss": 0.4512,
      "num_input_tokens_seen": 4987024,
      "step": 7635
    },
    {
      "epoch": 4.0041928721174,
      "grad_norm": 0.6395808458328247,
      "learning_rate": 4.8486835645904324e-05,
      "loss": 0.4263,
      "num_input_tokens_seen": 4989552,
      "step": 7640
    },
    {
      "epoch": 4.006813417190775,
      "grad_norm": 0.48461806774139404,
      "learning_rate": 4.84829155544286e-05,
      "loss": 0.684,
      "num_input_tokens_seen": 4992080,
      "step": 7645
    },
    {
      "epoch": 4.009433962264151,
      "grad_norm": 0.8604150414466858,
      "learning_rate": 4.847899055059484e-05,
      "loss": 0.6782,
      "num_input_tokens_seen": 4994512,
      "step": 7650
    },
    {
      "epoch": 4.012054507337526,
      "grad_norm": 0.3403312861919403,
      "learning_rate": 4.847506063522412e-05,
      "loss": 0.3104,
      "num_input_tokens_seen": 4998000,
      "step": 7655
    },
    {
      "epoch": 4.014675052410902,
      "grad_norm": 0.5854839086532593,
      "learning_rate": 4.847112580913854e-05,
      "loss": 0.4604,
      "num_input_tokens_seen": 5002000,
      "step": 7660
    },
    {
      "epoch": 4.017295597484277,
      "grad_norm": 0.4034757614135742,
      "learning_rate": 4.846718607316121e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 5005648,
      "step": 7665
    },
    {
      "epoch": 4.019916142557652,
      "grad_norm": 0.5856778621673584,
      "learning_rate": 4.846324142811628e-05,
      "loss": 0.6162,
      "num_input_tokens_seen": 5009104,
      "step": 7670
    },
    {
      "epoch": 4.022536687631027,
      "grad_norm": 0.3265402317047119,
      "learning_rate": 4.8459291874828924e-05,
      "loss": 0.3871,
      "num_input_tokens_seen": 5012304,
      "step": 7675
    },
    {
      "epoch": 4.0251572327044025,
      "grad_norm": 1.499530553817749,
      "learning_rate": 4.845533741412534e-05,
      "loss": 0.5256,
      "num_input_tokens_seen": 5014608,
      "step": 7680
    },
    {
      "epoch": 4.027777777777778,
      "grad_norm": 0.8615911602973938,
      "learning_rate": 4.845137804683277e-05,
      "loss": 0.5097,
      "num_input_tokens_seen": 5017776,
      "step": 7685
    },
    {
      "epoch": 4.030398322851153,
      "grad_norm": 0.4870012402534485,
      "learning_rate": 4.844741377377946e-05,
      "loss": 0.5973,
      "num_input_tokens_seen": 5020432,
      "step": 7690
    },
    {
      "epoch": 4.033018867924528,
      "grad_norm": 0.9556722640991211,
      "learning_rate": 4.844344459579469e-05,
      "loss": 0.3415,
      "num_input_tokens_seen": 5023824,
      "step": 7695
    },
    {
      "epoch": 4.035639412997903,
      "grad_norm": 0.32686272263526917,
      "learning_rate": 4.8439470513708766e-05,
      "loss": 0.517,
      "num_input_tokens_seen": 5027344,
      "step": 7700
    },
    {
      "epoch": 4.038259958071279,
      "grad_norm": 0.7688533067703247,
      "learning_rate": 4.8435491528353026e-05,
      "loss": 0.5076,
      "num_input_tokens_seen": 5030928,
      "step": 7705
    },
    {
      "epoch": 4.040880503144654,
      "grad_norm": 0.309280663728714,
      "learning_rate": 4.843150764055982e-05,
      "loss": 0.3892,
      "num_input_tokens_seen": 5034320,
      "step": 7710
    },
    {
      "epoch": 4.04350104821803,
      "grad_norm": 0.2523050308227539,
      "learning_rate": 4.842751885116254e-05,
      "loss": 0.4466,
      "num_input_tokens_seen": 5037456,
      "step": 7715
    },
    {
      "epoch": 4.046121593291405,
      "grad_norm": 0.3144902288913727,
      "learning_rate": 4.842352516099559e-05,
      "loss": 0.421,
      "num_input_tokens_seen": 5040784,
      "step": 7720
    },
    {
      "epoch": 4.04874213836478,
      "grad_norm": 0.6082159876823425,
      "learning_rate": 4.84195265708944e-05,
      "loss": 0.5854,
      "num_input_tokens_seen": 5044176,
      "step": 7725
    },
    {
      "epoch": 4.051362683438155,
      "grad_norm": 0.3990567624568939,
      "learning_rate": 4.841552308169545e-05,
      "loss": 0.5754,
      "num_input_tokens_seen": 5047376,
      "step": 7730
    },
    {
      "epoch": 4.05398322851153,
      "grad_norm": 0.363036572933197,
      "learning_rate": 4.841151469423619e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 5053712,
      "step": 7735
    },
    {
      "epoch": 4.056603773584905,
      "grad_norm": 0.6974440813064575,
      "learning_rate": 4.840750140935516e-05,
      "loss": 0.4808,
      "num_input_tokens_seen": 5057968,
      "step": 7740
    },
    {
      "epoch": 4.059224318658281,
      "grad_norm": 1.006972312927246,
      "learning_rate": 4.840348322789189e-05,
      "loss": 0.662,
      "num_input_tokens_seen": 5061328,
      "step": 7745
    },
    {
      "epoch": 4.061844863731656,
      "grad_norm": 0.5949901342391968,
      "learning_rate": 4.839946015068692e-05,
      "loss": 0.5127,
      "num_input_tokens_seen": 5064464,
      "step": 7750
    },
    {
      "epoch": 4.064465408805032,
      "grad_norm": 0.4232551157474518,
      "learning_rate": 4.839543217858184e-05,
      "loss": 0.4174,
      "num_input_tokens_seen": 5067344,
      "step": 7755
    },
    {
      "epoch": 4.067085953878407,
      "grad_norm": 0.3870261609554291,
      "learning_rate": 4.839139931241926e-05,
      "loss": 0.4603,
      "num_input_tokens_seen": 5070512,
      "step": 7760
    },
    {
      "epoch": 4.069706498951782,
      "grad_norm": 0.7966895699501038,
      "learning_rate": 4.838736155304281e-05,
      "loss": 0.3989,
      "num_input_tokens_seen": 5073488,
      "step": 7765
    },
    {
      "epoch": 4.072327044025157,
      "grad_norm": 0.5469701886177063,
      "learning_rate": 4.8383318901297135e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 5079984,
      "step": 7770
    },
    {
      "epoch": 4.0749475890985325,
      "grad_norm": 0.29883310198783875,
      "learning_rate": 4.837927135802792e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 5083056,
      "step": 7775
    },
    {
      "epoch": 4.077568134171908,
      "grad_norm": 0.634278416633606,
      "learning_rate": 4.8375218924081856e-05,
      "loss": 0.5332,
      "num_input_tokens_seen": 5085904,
      "step": 7780
    },
    {
      "epoch": 4.080188679245283,
      "grad_norm": 0.5078383684158325,
      "learning_rate": 4.837116160030668e-05,
      "loss": 0.4613,
      "num_input_tokens_seen": 5088368,
      "step": 7785
    },
    {
      "epoch": 4.082809224318658,
      "grad_norm": 0.3597625494003296,
      "learning_rate": 4.8367099387551126e-05,
      "loss": 0.4324,
      "num_input_tokens_seen": 5091088,
      "step": 7790
    },
    {
      "epoch": 4.085429769392033,
      "grad_norm": 0.31145745515823364,
      "learning_rate": 4.8363032286664964e-05,
      "loss": 0.4994,
      "num_input_tokens_seen": 5094960,
      "step": 7795
    },
    {
      "epoch": 4.088050314465409,
      "grad_norm": 0.415715754032135,
      "learning_rate": 4.835896029849899e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 5098352,
      "step": 7800
    },
    {
      "epoch": 4.090670859538784,
      "grad_norm": 1.084348201751709,
      "learning_rate": 4.835488342390502e-05,
      "loss": 0.4973,
      "num_input_tokens_seen": 5102416,
      "step": 7805
    },
    {
      "epoch": 4.09329140461216,
      "grad_norm": 0.3489917516708374,
      "learning_rate": 4.835080166373587e-05,
      "loss": 0.6054,
      "num_input_tokens_seen": 5105072,
      "step": 7810
    },
    {
      "epoch": 4.095911949685535,
      "grad_norm": 0.8165751695632935,
      "learning_rate": 4.834671501884542e-05,
      "loss": 0.5197,
      "num_input_tokens_seen": 5108528,
      "step": 7815
    },
    {
      "epoch": 4.09853249475891,
      "grad_norm": 0.41540592908859253,
      "learning_rate": 4.8342623490088543e-05,
      "loss": 0.5979,
      "num_input_tokens_seen": 5112176,
      "step": 7820
    },
    {
      "epoch": 4.101153039832285,
      "grad_norm": 0.263367235660553,
      "learning_rate": 4.833852707832114e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 5116848,
      "step": 7825
    },
    {
      "epoch": 4.10377358490566,
      "grad_norm": 0.31716325879096985,
      "learning_rate": 4.833442578440013e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 5120464,
      "step": 7830
    },
    {
      "epoch": 4.106394129979035,
      "grad_norm": 0.3138660192489624,
      "learning_rate": 4.833031960918346e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 5123312,
      "step": 7835
    },
    {
      "epoch": 4.109014675052411,
      "grad_norm": 0.492435485124588,
      "learning_rate": 4.8326208553530095e-05,
      "loss": 0.4682,
      "num_input_tokens_seen": 5127088,
      "step": 7840
    },
    {
      "epoch": 4.111635220125786,
      "grad_norm": 0.2893815338611603,
      "learning_rate": 4.832209261830002e-05,
      "loss": 0.4647,
      "num_input_tokens_seen": 5130384,
      "step": 7845
    },
    {
      "epoch": 4.114255765199162,
      "grad_norm": 0.3600878417491913,
      "learning_rate": 4.831797180435423e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 5134288,
      "step": 7850
    },
    {
      "epoch": 4.116876310272537,
      "grad_norm": 1.52946138381958,
      "learning_rate": 4.831384611255478e-05,
      "loss": 0.6091,
      "num_input_tokens_seen": 5137392,
      "step": 7855
    },
    {
      "epoch": 4.119496855345912,
      "grad_norm": 0.6285409927368164,
      "learning_rate": 4.8309715543764697e-05,
      "loss": 0.4728,
      "num_input_tokens_seen": 5139824,
      "step": 7860
    },
    {
      "epoch": 4.122117400419287,
      "grad_norm": 0.3139587640762329,
      "learning_rate": 4.8305580098848046e-05,
      "loss": 0.5585,
      "num_input_tokens_seen": 5143280,
      "step": 7865
    },
    {
      "epoch": 4.1247379454926625,
      "grad_norm": 0.5127859115600586,
      "learning_rate": 4.830143977866993e-05,
      "loss": 0.552,
      "num_input_tokens_seen": 5145584,
      "step": 7870
    },
    {
      "epoch": 4.127358490566038,
      "grad_norm": 0.4299357235431671,
      "learning_rate": 4.829729458409645e-05,
      "loss": 0.4399,
      "num_input_tokens_seen": 5148240,
      "step": 7875
    },
    {
      "epoch": 4.129979035639413,
      "grad_norm": 0.6511682868003845,
      "learning_rate": 4.8293144515994725e-05,
      "loss": 0.5753,
      "num_input_tokens_seen": 5151632,
      "step": 7880
    },
    {
      "epoch": 4.132599580712788,
      "grad_norm": 1.2386778593063354,
      "learning_rate": 4.828898957523292e-05,
      "loss": 0.5663,
      "num_input_tokens_seen": 5154576,
      "step": 7885
    },
    {
      "epoch": 4.135220125786163,
      "grad_norm": 0.43895643949508667,
      "learning_rate": 4.8284829762680183e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 5158224,
      "step": 7890
    },
    {
      "epoch": 4.137840670859539,
      "grad_norm": 0.3905109763145447,
      "learning_rate": 4.828066507920671e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 5161360,
      "step": 7895
    },
    {
      "epoch": 4.140461215932914,
      "grad_norm": 0.573762059211731,
      "learning_rate": 4.82764955256837e-05,
      "loss": 0.3935,
      "num_input_tokens_seen": 5164656,
      "step": 7900
    },
    {
      "epoch": 4.1430817610062896,
      "grad_norm": 0.41690415143966675,
      "learning_rate": 4.827232110298338e-05,
      "loss": 0.4641,
      "num_input_tokens_seen": 5168336,
      "step": 7905
    },
    {
      "epoch": 4.145702306079665,
      "grad_norm": 0.46584993600845337,
      "learning_rate": 4.8268141811978995e-05,
      "loss": 0.4956,
      "num_input_tokens_seen": 5170736,
      "step": 7910
    },
    {
      "epoch": 4.14832285115304,
      "grad_norm": 0.41016507148742676,
      "learning_rate": 4.826395765354479e-05,
      "loss": 0.2917,
      "num_input_tokens_seen": 5173424,
      "step": 7915
    },
    {
      "epoch": 4.150943396226415,
      "grad_norm": 0.4624747633934021,
      "learning_rate": 4.825976862855606e-05,
      "loss": 0.3656,
      "num_input_tokens_seen": 5175856,
      "step": 7920
    },
    {
      "epoch": 4.15356394129979,
      "grad_norm": 0.5386263132095337,
      "learning_rate": 4.8255574737889094e-05,
      "loss": 0.4283,
      "num_input_tokens_seen": 5178672,
      "step": 7925
    },
    {
      "epoch": 4.156184486373165,
      "grad_norm": 0.520839512348175,
      "learning_rate": 4.82513759824212e-05,
      "loss": 0.4513,
      "num_input_tokens_seen": 5184496,
      "step": 7930
    },
    {
      "epoch": 4.158805031446541,
      "grad_norm": 0.22899262607097626,
      "learning_rate": 4.824717236303072e-05,
      "loss": 0.5237,
      "num_input_tokens_seen": 5188144,
      "step": 7935
    },
    {
      "epoch": 4.161425576519916,
      "grad_norm": 0.7930534482002258,
      "learning_rate": 4.8242963880597005e-05,
      "loss": 0.4646,
      "num_input_tokens_seen": 5192112,
      "step": 7940
    },
    {
      "epoch": 4.164046121593292,
      "grad_norm": 0.5983563661575317,
      "learning_rate": 4.823875053600041e-05,
      "loss": 0.514,
      "num_input_tokens_seen": 5194320,
      "step": 7945
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 0.9421365261077881,
      "learning_rate": 4.823453233012233e-05,
      "loss": 0.4988,
      "num_input_tokens_seen": 5197456,
      "step": 7950
    },
    {
      "epoch": 4.169287211740042,
      "grad_norm": 0.3391250967979431,
      "learning_rate": 4.8230309263845155e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 5199792,
      "step": 7955
    },
    {
      "epoch": 4.171907756813417,
      "grad_norm": 0.5240896344184875,
      "learning_rate": 4.822608133805231e-05,
      "loss": 0.6097,
      "num_input_tokens_seen": 5202672,
      "step": 7960
    },
    {
      "epoch": 4.1745283018867925,
      "grad_norm": 0.7365862131118774,
      "learning_rate": 4.8221848553628226e-05,
      "loss": 0.5835,
      "num_input_tokens_seen": 5205808,
      "step": 7965
    },
    {
      "epoch": 4.177148846960168,
      "grad_norm": 0.6310661435127258,
      "learning_rate": 4.8217610911458346e-05,
      "loss": 0.3461,
      "num_input_tokens_seen": 5208176,
      "step": 7970
    },
    {
      "epoch": 4.179769392033543,
      "grad_norm": 0.3390653729438782,
      "learning_rate": 4.821336841242915e-05,
      "loss": 0.4785,
      "num_input_tokens_seen": 5215376,
      "step": 7975
    },
    {
      "epoch": 4.182389937106918,
      "grad_norm": 0.3965795934200287,
      "learning_rate": 4.820912105742811e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 5218064,
      "step": 7980
    },
    {
      "epoch": 4.185010482180293,
      "grad_norm": 0.8053150177001953,
      "learning_rate": 4.8204868847343736e-05,
      "loss": 0.5091,
      "num_input_tokens_seen": 5221328,
      "step": 7985
    },
    {
      "epoch": 4.187631027253669,
      "grad_norm": 0.4046705663204193,
      "learning_rate": 4.820061178306553e-05,
      "loss": 0.483,
      "num_input_tokens_seen": 5224432,
      "step": 7990
    },
    {
      "epoch": 4.190251572327044,
      "grad_norm": 0.5491102933883667,
      "learning_rate": 4.819634986548403e-05,
      "loss": 0.4869,
      "num_input_tokens_seen": 5226896,
      "step": 7995
    },
    {
      "epoch": 4.1928721174004195,
      "grad_norm": 0.4248526692390442,
      "learning_rate": 4.819208309549077e-05,
      "loss": 0.4619,
      "num_input_tokens_seen": 5231056,
      "step": 8000
    },
    {
      "epoch": 4.195492662473795,
      "grad_norm": 0.5593143701553345,
      "learning_rate": 4.8187811473978316e-05,
      "loss": 0.398,
      "num_input_tokens_seen": 5234480,
      "step": 8005
    },
    {
      "epoch": 4.19811320754717,
      "grad_norm": 2.0783469676971436,
      "learning_rate": 4.818353500184025e-05,
      "loss": 0.6436,
      "num_input_tokens_seen": 5236816,
      "step": 8010
    },
    {
      "epoch": 4.200733752620545,
      "grad_norm": 1.4125639200210571,
      "learning_rate": 4.817925367997115e-05,
      "loss": 0.4713,
      "num_input_tokens_seen": 5239600,
      "step": 8015
    },
    {
      "epoch": 4.20335429769392,
      "grad_norm": 0.4748423993587494,
      "learning_rate": 4.8174967509266615e-05,
      "loss": 0.5083,
      "num_input_tokens_seen": 5242896,
      "step": 8020
    },
    {
      "epoch": 4.205974842767295,
      "grad_norm": 0.5393669009208679,
      "learning_rate": 4.817067649062328e-05,
      "loss": 0.5543,
      "num_input_tokens_seen": 5245904,
      "step": 8025
    },
    {
      "epoch": 4.2085953878406706,
      "grad_norm": 0.5540959239006042,
      "learning_rate": 4.816638062493876e-05,
      "loss": 0.5729,
      "num_input_tokens_seen": 5249968,
      "step": 8030
    },
    {
      "epoch": 4.211215932914046,
      "grad_norm": 0.4617502689361572,
      "learning_rate": 4.8162079913111716e-05,
      "loss": 0.4402,
      "num_input_tokens_seen": 5253680,
      "step": 8035
    },
    {
      "epoch": 4.213836477987422,
      "grad_norm": 0.5904155373573303,
      "learning_rate": 4.81577743560418e-05,
      "loss": 0.6512,
      "num_input_tokens_seen": 5256144,
      "step": 8040
    },
    {
      "epoch": 4.216457023060797,
      "grad_norm": 0.366159588098526,
      "learning_rate": 4.8153463954629684e-05,
      "loss": 0.3756,
      "num_input_tokens_seen": 5261456,
      "step": 8045
    },
    {
      "epoch": 4.219077568134172,
      "grad_norm": 0.4826522767543793,
      "learning_rate": 4.814914870977706e-05,
      "loss": 0.5582,
      "num_input_tokens_seen": 5264688,
      "step": 8050
    },
    {
      "epoch": 4.221698113207547,
      "grad_norm": 0.38004982471466064,
      "learning_rate": 4.8144828622386627e-05,
      "loss": 0.6416,
      "num_input_tokens_seen": 5267184,
      "step": 8055
    },
    {
      "epoch": 4.2243186582809225,
      "grad_norm": 0.9289650917053223,
      "learning_rate": 4.8140503693362095e-05,
      "loss": 0.5253,
      "num_input_tokens_seen": 5269776,
      "step": 8060
    },
    {
      "epoch": 4.226939203354298,
      "grad_norm": 0.30018359422683716,
      "learning_rate": 4.813617392360818e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 5273296,
      "step": 8065
    },
    {
      "epoch": 4.229559748427673,
      "grad_norm": 0.5184326171875,
      "learning_rate": 4.813183931403065e-05,
      "loss": 0.4708,
      "num_input_tokens_seen": 5276432,
      "step": 8070
    },
    {
      "epoch": 4.232180293501048,
      "grad_norm": 0.45933371782302856,
      "learning_rate": 4.8127499865536226e-05,
      "loss": 0.4042,
      "num_input_tokens_seen": 5279696,
      "step": 8075
    },
    {
      "epoch": 4.234800838574423,
      "grad_norm": 0.44512370228767395,
      "learning_rate": 4.812315557903269e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 5283312,
      "step": 8080
    },
    {
      "epoch": 4.237421383647799,
      "grad_norm": 0.5568438172340393,
      "learning_rate": 4.8118806455428806e-05,
      "loss": 0.3833,
      "num_input_tokens_seen": 5285712,
      "step": 8085
    },
    {
      "epoch": 4.240041928721174,
      "grad_norm": 0.33500179648399353,
      "learning_rate": 4.811445249563437e-05,
      "loss": 0.5157,
      "num_input_tokens_seen": 5289136,
      "step": 8090
    },
    {
      "epoch": 4.2426624737945495,
      "grad_norm": 0.41948065161705017,
      "learning_rate": 4.811009370056017e-05,
      "loss": 0.3127,
      "num_input_tokens_seen": 5292528,
      "step": 8095
    },
    {
      "epoch": 4.245283018867925,
      "grad_norm": 0.3484889268875122,
      "learning_rate": 4.8105730071118016e-05,
      "loss": 0.529,
      "num_input_tokens_seen": 5295984,
      "step": 8100
    },
    {
      "epoch": 4.2479035639413,
      "grad_norm": 0.4770960211753845,
      "learning_rate": 4.810136160822074e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 5299120,
      "step": 8105
    },
    {
      "epoch": 4.250524109014675,
      "grad_norm": 0.4611732065677643,
      "learning_rate": 4.8096988312782174e-05,
      "loss": 0.4208,
      "num_input_tokens_seen": 5302192,
      "step": 8110
    },
    {
      "epoch": 4.25314465408805,
      "grad_norm": 0.8288554549217224,
      "learning_rate": 4.809261018571715e-05,
      "loss": 0.8122,
      "num_input_tokens_seen": 5304880,
      "step": 8115
    },
    {
      "epoch": 4.255765199161425,
      "grad_norm": 0.414760023355484,
      "learning_rate": 4.808822722794153e-05,
      "loss": 0.5485,
      "num_input_tokens_seen": 5308304,
      "step": 8120
    },
    {
      "epoch": 4.2583857442348005,
      "grad_norm": 0.5442935824394226,
      "learning_rate": 4.8083839440372194e-05,
      "loss": 0.5448,
      "num_input_tokens_seen": 5310960,
      "step": 8125
    },
    {
      "epoch": 4.261006289308176,
      "grad_norm": 0.4062289893627167,
      "learning_rate": 4.8079446823926984e-05,
      "loss": 0.4466,
      "num_input_tokens_seen": 5313264,
      "step": 8130
    },
    {
      "epoch": 4.263626834381552,
      "grad_norm": 0.44661784172058105,
      "learning_rate": 4.8075049379524805e-05,
      "loss": 0.5642,
      "num_input_tokens_seen": 5315600,
      "step": 8135
    },
    {
      "epoch": 4.266247379454927,
      "grad_norm": 0.436716765165329,
      "learning_rate": 4.807064710808555e-05,
      "loss": 0.3765,
      "num_input_tokens_seen": 5319024,
      "step": 8140
    },
    {
      "epoch": 4.268867924528302,
      "grad_norm": 0.5830195546150208,
      "learning_rate": 4.806624001053013e-05,
      "loss": 0.4708,
      "num_input_tokens_seen": 5322160,
      "step": 8145
    },
    {
      "epoch": 4.271488469601677,
      "grad_norm": 0.3806711733341217,
      "learning_rate": 4.8061828087780455e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 5327152,
      "step": 8150
    },
    {
      "epoch": 4.274109014675052,
      "grad_norm": 0.482404887676239,
      "learning_rate": 4.805741134075944e-05,
      "loss": 0.3674,
      "num_input_tokens_seen": 5330256,
      "step": 8155
    },
    {
      "epoch": 4.276729559748428,
      "grad_norm": 0.703994631767273,
      "learning_rate": 4.8052989770391026e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 5333392,
      "step": 8160
    },
    {
      "epoch": 4.279350104821803,
      "grad_norm": 0.4567719101905823,
      "learning_rate": 4.804856337760016e-05,
      "loss": 0.4602,
      "num_input_tokens_seen": 5336720,
      "step": 8165
    },
    {
      "epoch": 4.281970649895178,
      "grad_norm": 0.33088797330856323,
      "learning_rate": 4.804413216331278e-05,
      "loss": 0.3079,
      "num_input_tokens_seen": 5339696,
      "step": 8170
    },
    {
      "epoch": 4.284591194968553,
      "grad_norm": 0.5295208692550659,
      "learning_rate": 4.8039696128455856e-05,
      "loss": 0.5626,
      "num_input_tokens_seen": 5342896,
      "step": 8175
    },
    {
      "epoch": 4.287211740041929,
      "grad_norm": 0.7422252297401428,
      "learning_rate": 4.803525527395736e-05,
      "loss": 0.5101,
      "num_input_tokens_seen": 5345392,
      "step": 8180
    },
    {
      "epoch": 4.289832285115304,
      "grad_norm": 0.5236266851425171,
      "learning_rate": 4.803080960074625e-05,
      "loss": 0.444,
      "num_input_tokens_seen": 5349392,
      "step": 8185
    },
    {
      "epoch": 4.2924528301886795,
      "grad_norm": 0.49119263887405396,
      "learning_rate": 4.8026359109752525e-05,
      "loss": 0.4598,
      "num_input_tokens_seen": 5353008,
      "step": 8190
    },
    {
      "epoch": 4.295073375262055,
      "grad_norm": 0.7283792495727539,
      "learning_rate": 4.8021903801907175e-05,
      "loss": 0.7337,
      "num_input_tokens_seen": 5356464,
      "step": 8195
    },
    {
      "epoch": 4.29769392033543,
      "grad_norm": 0.3800901472568512,
      "learning_rate": 4.801744367814219e-05,
      "loss": 0.4098,
      "num_input_tokens_seen": 5359440,
      "step": 8200
    },
    {
      "epoch": 4.300314465408805,
      "grad_norm": 0.6062473654747009,
      "learning_rate": 4.801297873939059e-05,
      "loss": 0.6136,
      "num_input_tokens_seen": 5361968,
      "step": 8205
    },
    {
      "epoch": 4.30293501048218,
      "grad_norm": 0.3484911322593689,
      "learning_rate": 4.8008508986586384e-05,
      "loss": 0.6534,
      "num_input_tokens_seen": 5365744,
      "step": 8210
    },
    {
      "epoch": 4.305555555555555,
      "grad_norm": 0.5291152596473694,
      "learning_rate": 4.800403442066459e-05,
      "loss": 0.3821,
      "num_input_tokens_seen": 5368720,
      "step": 8215
    },
    {
      "epoch": 4.3081761006289305,
      "grad_norm": 1.2023848295211792,
      "learning_rate": 4.799955504256125e-05,
      "loss": 0.5193,
      "num_input_tokens_seen": 5371408,
      "step": 8220
    },
    {
      "epoch": 4.310796645702306,
      "grad_norm": 0.8944399356842041,
      "learning_rate": 4.7995070853213376e-05,
      "loss": 0.5632,
      "num_input_tokens_seen": 5374256,
      "step": 8225
    },
    {
      "epoch": 4.313417190775682,
      "grad_norm": 0.8695749640464783,
      "learning_rate": 4.799058185355903e-05,
      "loss": 0.5595,
      "num_input_tokens_seen": 5376624,
      "step": 8230
    },
    {
      "epoch": 4.316037735849057,
      "grad_norm": 0.6228213310241699,
      "learning_rate": 4.7986088044537235e-05,
      "loss": 0.4827,
      "num_input_tokens_seen": 5380048,
      "step": 8235
    },
    {
      "epoch": 4.318658280922432,
      "grad_norm": 0.4538964629173279,
      "learning_rate": 4.798158942708807e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 5382416,
      "step": 8240
    },
    {
      "epoch": 4.321278825995807,
      "grad_norm": 0.37981319427490234,
      "learning_rate": 4.797708600215258e-05,
      "loss": 0.6299,
      "num_input_tokens_seen": 5385520,
      "step": 8245
    },
    {
      "epoch": 4.323899371069182,
      "grad_norm": 0.3733126223087311,
      "learning_rate": 4.7972577770672845e-05,
      "loss": 0.6874,
      "num_input_tokens_seen": 5388368,
      "step": 8250
    },
    {
      "epoch": 4.326519916142558,
      "grad_norm": 0.4031426012516022,
      "learning_rate": 4.796806473359192e-05,
      "loss": 0.4215,
      "num_input_tokens_seen": 5391920,
      "step": 8255
    },
    {
      "epoch": 4.329140461215933,
      "grad_norm": 0.4068942964076996,
      "learning_rate": 4.796354689185389e-05,
      "loss": 0.3731,
      "num_input_tokens_seen": 5394864,
      "step": 8260
    },
    {
      "epoch": 4.331761006289308,
      "grad_norm": 0.4146069884300232,
      "learning_rate": 4.795902424640383e-05,
      "loss": 0.4541,
      "num_input_tokens_seen": 5398448,
      "step": 8265
    },
    {
      "epoch": 4.334381551362683,
      "grad_norm": 0.47927147150039673,
      "learning_rate": 4.795449679818782e-05,
      "loss": 0.427,
      "num_input_tokens_seen": 5401904,
      "step": 8270
    },
    {
      "epoch": 4.337002096436059,
      "grad_norm": 0.48666098713874817,
      "learning_rate": 4.794996454815297e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 5404432,
      "step": 8275
    },
    {
      "epoch": 4.339622641509434,
      "grad_norm": 0.685822069644928,
      "learning_rate": 4.7945427497247355e-05,
      "loss": 0.4876,
      "num_input_tokens_seen": 5407376,
      "step": 8280
    },
    {
      "epoch": 4.3422431865828095,
      "grad_norm": 0.9782160520553589,
      "learning_rate": 4.7940885646420084e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 5410768,
      "step": 8285
    },
    {
      "epoch": 4.344863731656185,
      "grad_norm": 0.3160076439380646,
      "learning_rate": 4.7936338996621264e-05,
      "loss": 0.5265,
      "num_input_tokens_seen": 5414256,
      "step": 8290
    },
    {
      "epoch": 4.34748427672956,
      "grad_norm": 0.543267011642456,
      "learning_rate": 4.7931787548801996e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 5417296,
      "step": 8295
    },
    {
      "epoch": 4.350104821802935,
      "grad_norm": 0.6403440237045288,
      "learning_rate": 4.7927231303914394e-05,
      "loss": 0.4819,
      "num_input_tokens_seen": 5420560,
      "step": 8300
    },
    {
      "epoch": 4.35272536687631,
      "grad_norm": 0.663875937461853,
      "learning_rate": 4.7922670262911565e-05,
      "loss": 0.4433,
      "num_input_tokens_seen": 5424496,
      "step": 8305
    },
    {
      "epoch": 4.355345911949685,
      "grad_norm": 0.3191729784011841,
      "learning_rate": 4.791810442674764e-05,
      "loss": 0.4779,
      "num_input_tokens_seen": 5428944,
      "step": 8310
    },
    {
      "epoch": 4.3579664570230605,
      "grad_norm": 0.2938316762447357,
      "learning_rate": 4.791353379637773e-05,
      "loss": 0.5309,
      "num_input_tokens_seen": 5432624,
      "step": 8315
    },
    {
      "epoch": 4.360587002096436,
      "grad_norm": 0.7051478028297424,
      "learning_rate": 4.790895837275796e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 5435152,
      "step": 8320
    },
    {
      "epoch": 4.363207547169811,
      "grad_norm": 0.6256168484687805,
      "learning_rate": 4.790437815684545e-05,
      "loss": 0.511,
      "num_input_tokens_seen": 5438640,
      "step": 8325
    },
    {
      "epoch": 4.365828092243187,
      "grad_norm": 0.4634554088115692,
      "learning_rate": 4.789979314959835e-05,
      "loss": 0.4397,
      "num_input_tokens_seen": 5441264,
      "step": 8330
    },
    {
      "epoch": 4.368448637316562,
      "grad_norm": 0.29744720458984375,
      "learning_rate": 4.789520335197577e-05,
      "loss": 0.4361,
      "num_input_tokens_seen": 5444752,
      "step": 8335
    },
    {
      "epoch": 4.371069182389937,
      "grad_norm": 0.6470555663108826,
      "learning_rate": 4.789060876493785e-05,
      "loss": 0.4912,
      "num_input_tokens_seen": 5447632,
      "step": 8340
    },
    {
      "epoch": 4.373689727463312,
      "grad_norm": 0.47190049290657043,
      "learning_rate": 4.788600938944573e-05,
      "loss": 0.6039,
      "num_input_tokens_seen": 5451344,
      "step": 8345
    },
    {
      "epoch": 4.376310272536688,
      "grad_norm": 0.4877948462963104,
      "learning_rate": 4.788140522646155e-05,
      "loss": 0.5634,
      "num_input_tokens_seen": 5454736,
      "step": 8350
    },
    {
      "epoch": 4.378930817610063,
      "grad_norm": 0.31084856390953064,
      "learning_rate": 4.787679627694843e-05,
      "loss": 0.5143,
      "num_input_tokens_seen": 5458096,
      "step": 8355
    },
    {
      "epoch": 4.381551362683438,
      "grad_norm": 0.44352987408638,
      "learning_rate": 4.787218254187053e-05,
      "loss": 0.481,
      "num_input_tokens_seen": 5460560,
      "step": 8360
    },
    {
      "epoch": 4.384171907756813,
      "grad_norm": 0.5894736051559448,
      "learning_rate": 4.7867564022192965e-05,
      "loss": 0.4558,
      "num_input_tokens_seen": 5463792,
      "step": 8365
    },
    {
      "epoch": 4.386792452830189,
      "grad_norm": 0.3868297040462494,
      "learning_rate": 4.786294071888191e-05,
      "loss": 0.4488,
      "num_input_tokens_seen": 5467216,
      "step": 8370
    },
    {
      "epoch": 4.389412997903564,
      "grad_norm": 0.2500629723072052,
      "learning_rate": 4.785831263290449e-05,
      "loss": 0.4526,
      "num_input_tokens_seen": 5470320,
      "step": 8375
    },
    {
      "epoch": 4.3920335429769395,
      "grad_norm": 0.3803715109825134,
      "learning_rate": 4.7853679765228845e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 5473232,
      "step": 8380
    },
    {
      "epoch": 4.394654088050315,
      "grad_norm": 0.5931675434112549,
      "learning_rate": 4.7849042116824116e-05,
      "loss": 0.6068,
      "num_input_tokens_seen": 5476432,
      "step": 8385
    },
    {
      "epoch": 4.39727463312369,
      "grad_norm": 0.7734373211860657,
      "learning_rate": 4.7844399688660456e-05,
      "loss": 0.5658,
      "num_input_tokens_seen": 5478928,
      "step": 8390
    },
    {
      "epoch": 4.399895178197065,
      "grad_norm": 0.3613705337047577,
      "learning_rate": 4.783975248170901e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 5481776,
      "step": 8395
    },
    {
      "epoch": 4.40251572327044,
      "grad_norm": 0.32045987248420715,
      "learning_rate": 4.783510049694191e-05,
      "loss": 0.4712,
      "num_input_tokens_seen": 5485008,
      "step": 8400
    },
    {
      "epoch": 4.405136268343815,
      "grad_norm": 0.26700326800346375,
      "learning_rate": 4.78304437353323e-05,
      "loss": 0.4925,
      "num_input_tokens_seen": 5492496,
      "step": 8405
    },
    {
      "epoch": 4.4077568134171905,
      "grad_norm": 0.5479003190994263,
      "learning_rate": 4.782578219785433e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 5495664,
      "step": 8410
    },
    {
      "epoch": 4.410377358490566,
      "grad_norm": 0.5552907586097717,
      "learning_rate": 4.782111588548312e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 5499792,
      "step": 8415
    },
    {
      "epoch": 4.412997903563941,
      "grad_norm": 0.3219716250896454,
      "learning_rate": 4.781644479919484e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 5502736,
      "step": 8420
    },
    {
      "epoch": 4.415618448637317,
      "grad_norm": 0.51189124584198,
      "learning_rate": 4.7811768939966605e-05,
      "loss": 0.4352,
      "num_input_tokens_seen": 5505328,
      "step": 8425
    },
    {
      "epoch": 4.418238993710692,
      "grad_norm": 0.600403904914856,
      "learning_rate": 4.780708830877656e-05,
      "loss": 0.506,
      "num_input_tokens_seen": 5507952,
      "step": 8430
    },
    {
      "epoch": 4.420859538784067,
      "grad_norm": 1.2580472230911255,
      "learning_rate": 4.780240290660384e-05,
      "loss": 0.3941,
      "num_input_tokens_seen": 5510544,
      "step": 8435
    },
    {
      "epoch": 4.423480083857442,
      "grad_norm": 0.48287034034729004,
      "learning_rate": 4.7797712734428575e-05,
      "loss": 0.553,
      "num_input_tokens_seen": 5512976,
      "step": 8440
    },
    {
      "epoch": 4.426100628930818,
      "grad_norm": 0.3479578197002411,
      "learning_rate": 4.779301779323189e-05,
      "loss": 0.5409,
      "num_input_tokens_seen": 5516496,
      "step": 8445
    },
    {
      "epoch": 4.428721174004193,
      "grad_norm": 0.4329839050769806,
      "learning_rate": 4.778831808399593e-05,
      "loss": 0.3667,
      "num_input_tokens_seen": 5520720,
      "step": 8450
    },
    {
      "epoch": 4.431341719077568,
      "grad_norm": 0.26972874999046326,
      "learning_rate": 4.77836136077038e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 5524336,
      "step": 8455
    },
    {
      "epoch": 4.433962264150943,
      "grad_norm": 0.23230619728565216,
      "learning_rate": 4.777890436533963e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 5528112,
      "step": 8460
    },
    {
      "epoch": 4.436582809224318,
      "grad_norm": 0.33414921164512634,
      "learning_rate": 4.777419035788855e-05,
      "loss": 0.529,
      "num_input_tokens_seen": 5531664,
      "step": 8465
    },
    {
      "epoch": 4.439203354297694,
      "grad_norm": 0.7671618461608887,
      "learning_rate": 4.776947158633667e-05,
      "loss": 0.4046,
      "num_input_tokens_seen": 5536368,
      "step": 8470
    },
    {
      "epoch": 4.4418238993710695,
      "grad_norm": 0.38046377897262573,
      "learning_rate": 4.7764748051671095e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 5539600,
      "step": 8475
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 0.3529159128665924,
      "learning_rate": 4.7760019754879946e-05,
      "loss": 0.4155,
      "num_input_tokens_seen": 5542544,
      "step": 8480
    },
    {
      "epoch": 4.44706498951782,
      "grad_norm": 0.3374272882938385,
      "learning_rate": 4.7755286696952324e-05,
      "loss": 0.4619,
      "num_input_tokens_seen": 5546704,
      "step": 8485
    },
    {
      "epoch": 4.449685534591195,
      "grad_norm": 0.34021738171577454,
      "learning_rate": 4.775054887887833e-05,
      "loss": 0.4168,
      "num_input_tokens_seen": 5549520,
      "step": 8490
    },
    {
      "epoch": 4.45230607966457,
      "grad_norm": 0.7263852953910828,
      "learning_rate": 4.774580630164906e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 5553040,
      "step": 8495
    },
    {
      "epoch": 4.454926624737945,
      "grad_norm": 0.3422749936580658,
      "learning_rate": 4.774105896625661e-05,
      "loss": 0.4959,
      "num_input_tokens_seen": 5556880,
      "step": 8500
    },
    {
      "epoch": 4.4575471698113205,
      "grad_norm": 0.4587191343307495,
      "learning_rate": 4.773630687369407e-05,
      "loss": 0.3916,
      "num_input_tokens_seen": 5559760,
      "step": 8505
    },
    {
      "epoch": 4.460167714884696,
      "grad_norm": 0.30281513929367065,
      "learning_rate": 4.773155002495552e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 5564688,
      "step": 8510
    },
    {
      "epoch": 4.462788259958071,
      "grad_norm": 0.6009988784790039,
      "learning_rate": 4.772678842103605e-05,
      "loss": 0.5254,
      "num_input_tokens_seen": 5568112,
      "step": 8515
    },
    {
      "epoch": 4.465408805031447,
      "grad_norm": 0.575054943561554,
      "learning_rate": 4.772202206293172e-05,
      "loss": 0.5807,
      "num_input_tokens_seen": 5570928,
      "step": 8520
    },
    {
      "epoch": 4.468029350104822,
      "grad_norm": 0.35117900371551514,
      "learning_rate": 4.7717250951639586e-05,
      "loss": 0.4784,
      "num_input_tokens_seen": 5574480,
      "step": 8525
    },
    {
      "epoch": 4.470649895178197,
      "grad_norm": 0.7208465933799744,
      "learning_rate": 4.771247508815774e-05,
      "loss": 0.4875,
      "num_input_tokens_seen": 5577296,
      "step": 8530
    },
    {
      "epoch": 4.473270440251572,
      "grad_norm": 0.2734467089176178,
      "learning_rate": 4.770769447348522e-05,
      "loss": 0.3772,
      "num_input_tokens_seen": 5580784,
      "step": 8535
    },
    {
      "epoch": 4.475890985324948,
      "grad_norm": 0.3554143011569977,
      "learning_rate": 4.7702909108622074e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 5583856,
      "step": 8540
    },
    {
      "epoch": 4.478511530398323,
      "grad_norm": 0.4457516372203827,
      "learning_rate": 4.7698118994569355e-05,
      "loss": 0.4661,
      "num_input_tokens_seen": 5586256,
      "step": 8545
    },
    {
      "epoch": 4.481132075471698,
      "grad_norm": 0.47251230478286743,
      "learning_rate": 4.769332413232909e-05,
      "loss": 0.4496,
      "num_input_tokens_seen": 5589200,
      "step": 8550
    },
    {
      "epoch": 4.483752620545073,
      "grad_norm": 0.37225258350372314,
      "learning_rate": 4.768852452290432e-05,
      "loss": 0.5467,
      "num_input_tokens_seen": 5592976,
      "step": 8555
    },
    {
      "epoch": 4.486373165618448,
      "grad_norm": 0.577662467956543,
      "learning_rate": 4.768372016729907e-05,
      "loss": 0.5139,
      "num_input_tokens_seen": 5595696,
      "step": 8560
    },
    {
      "epoch": 4.488993710691824,
      "grad_norm": 0.5083913207054138,
      "learning_rate": 4.767891106651834e-05,
      "loss": 0.5079,
      "num_input_tokens_seen": 5598288,
      "step": 8565
    },
    {
      "epoch": 4.4916142557651995,
      "grad_norm": 0.3951677680015564,
      "learning_rate": 4.767409722156816e-05,
      "loss": 0.5452,
      "num_input_tokens_seen": 5601296,
      "step": 8570
    },
    {
      "epoch": 4.494234800838575,
      "grad_norm": 1.2138384580612183,
      "learning_rate": 4.7669278633455515e-05,
      "loss": 0.3899,
      "num_input_tokens_seen": 5604752,
      "step": 8575
    },
    {
      "epoch": 4.49685534591195,
      "grad_norm": 0.4641764760017395,
      "learning_rate": 4.7664455303188404e-05,
      "loss": 0.5313,
      "num_input_tokens_seen": 5607824,
      "step": 8580
    },
    {
      "epoch": 4.499475890985325,
      "grad_norm": 0.4829716086387634,
      "learning_rate": 4.7659627231775815e-05,
      "loss": 0.514,
      "num_input_tokens_seen": 5611248,
      "step": 8585
    },
    {
      "epoch": 4.5020964360587,
      "grad_norm": 0.6227415800094604,
      "learning_rate": 4.765479442022773e-05,
      "loss": 0.4298,
      "num_input_tokens_seen": 5614192,
      "step": 8590
    },
    {
      "epoch": 4.504716981132075,
      "grad_norm": 0.5810482501983643,
      "learning_rate": 4.76499568695551e-05,
      "loss": 0.5377,
      "num_input_tokens_seen": 5616944,
      "step": 8595
    },
    {
      "epoch": 4.5073375262054505,
      "grad_norm": 0.6199681758880615,
      "learning_rate": 4.764511458076991e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 5619344,
      "step": 8600
    },
    {
      "epoch": 4.509958071278826,
      "grad_norm": 0.35374000668525696,
      "learning_rate": 4.7640267554885085e-05,
      "loss": 0.4079,
      "num_input_tokens_seen": 5622608,
      "step": 8605
    },
    {
      "epoch": 4.512578616352201,
      "grad_norm": 1.0341522693634033,
      "learning_rate": 4.763541579291459e-05,
      "loss": 0.4962,
      "num_input_tokens_seen": 5625296,
      "step": 8610
    },
    {
      "epoch": 4.515199161425577,
      "grad_norm": 0.46705126762390137,
      "learning_rate": 4.763055929587336e-05,
      "loss": 0.3825,
      "num_input_tokens_seen": 5628272,
      "step": 8615
    },
    {
      "epoch": 4.517819706498952,
      "grad_norm": 0.3375050127506256,
      "learning_rate": 4.7625698064777294e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 5632944,
      "step": 8620
    },
    {
      "epoch": 4.520440251572327,
      "grad_norm": 0.588157594203949,
      "learning_rate": 4.762083210064332e-05,
      "loss": 0.451,
      "num_input_tokens_seen": 5636016,
      "step": 8625
    },
    {
      "epoch": 4.523060796645702,
      "grad_norm": 0.6006048917770386,
      "learning_rate": 4.761596140448934e-05,
      "loss": 0.4177,
      "num_input_tokens_seen": 5642064,
      "step": 8630
    },
    {
      "epoch": 4.5256813417190775,
      "grad_norm": 0.47556933760643005,
      "learning_rate": 4.7611085977334256e-05,
      "loss": 0.5394,
      "num_input_tokens_seen": 5645072,
      "step": 8635
    },
    {
      "epoch": 4.528301886792453,
      "grad_norm": 0.36540377140045166,
      "learning_rate": 4.7606205820197946e-05,
      "loss": 0.5338,
      "num_input_tokens_seen": 5648304,
      "step": 8640
    },
    {
      "epoch": 4.530922431865828,
      "grad_norm": 0.4543743431568146,
      "learning_rate": 4.760132093410129e-05,
      "loss": 0.593,
      "num_input_tokens_seen": 5651120,
      "step": 8645
    },
    {
      "epoch": 4.533542976939203,
      "grad_norm": 1.0034527778625488,
      "learning_rate": 4.759643132006614e-05,
      "loss": 0.3986,
      "num_input_tokens_seen": 5653712,
      "step": 8650
    },
    {
      "epoch": 4.536163522012579,
      "grad_norm": 0.41829583048820496,
      "learning_rate": 4.759153697911535e-05,
      "loss": 0.4791,
      "num_input_tokens_seen": 5656624,
      "step": 8655
    },
    {
      "epoch": 4.538784067085954,
      "grad_norm": 0.4847331643104553,
      "learning_rate": 4.758663791227276e-05,
      "loss": 0.4308,
      "num_input_tokens_seen": 5658960,
      "step": 8660
    },
    {
      "epoch": 4.5414046121593294,
      "grad_norm": 0.5790786147117615,
      "learning_rate": 4.7581734120563206e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 5661872,
      "step": 8665
    },
    {
      "epoch": 4.544025157232705,
      "grad_norm": 0.34622910618782043,
      "learning_rate": 4.757682560501249e-05,
      "loss": 0.5741,
      "num_input_tokens_seen": 5664560,
      "step": 8670
    },
    {
      "epoch": 4.54664570230608,
      "grad_norm": 0.4264695644378662,
      "learning_rate": 4.757191236664744e-05,
      "loss": 0.4157,
      "num_input_tokens_seen": 5667664,
      "step": 8675
    },
    {
      "epoch": 4.549266247379455,
      "grad_norm": 0.5486508011817932,
      "learning_rate": 4.7566994406495825e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 5670128,
      "step": 8680
    },
    {
      "epoch": 4.55188679245283,
      "grad_norm": 0.4913181960582733,
      "learning_rate": 4.756207172558645e-05,
      "loss": 0.5485,
      "num_input_tokens_seen": 5673296,
      "step": 8685
    },
    {
      "epoch": 4.554507337526205,
      "grad_norm": 0.4657289981842041,
      "learning_rate": 4.7557144324949065e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 5675920,
      "step": 8690
    },
    {
      "epoch": 4.5571278825995805,
      "grad_norm": 1.0684328079223633,
      "learning_rate": 4.755221220561443e-05,
      "loss": 0.4397,
      "num_input_tokens_seen": 5678896,
      "step": 8695
    },
    {
      "epoch": 4.559748427672956,
      "grad_norm": 0.8690307140350342,
      "learning_rate": 4.7547275368614295e-05,
      "loss": 0.5211,
      "num_input_tokens_seen": 5681488,
      "step": 8700
    },
    {
      "epoch": 4.562368972746331,
      "grad_norm": 0.9456200003623962,
      "learning_rate": 4.754233381498139e-05,
      "loss": 0.5117,
      "num_input_tokens_seen": 5684432,
      "step": 8705
    },
    {
      "epoch": 4.564989517819707,
      "grad_norm": 0.37291207909584045,
      "learning_rate": 4.753738754574942e-05,
      "loss": 0.4719,
      "num_input_tokens_seen": 5688112,
      "step": 8710
    },
    {
      "epoch": 4.567610062893082,
      "grad_norm": 0.4282878637313843,
      "learning_rate": 4.75324365619531e-05,
      "loss": 0.5246,
      "num_input_tokens_seen": 5690416,
      "step": 8715
    },
    {
      "epoch": 4.570230607966457,
      "grad_norm": 0.43644818663597107,
      "learning_rate": 4.752748086462811e-05,
      "loss": 0.5341,
      "num_input_tokens_seen": 5693744,
      "step": 8720
    },
    {
      "epoch": 4.572851153039832,
      "grad_norm": 0.2803589105606079,
      "learning_rate": 4.752252045481113e-05,
      "loss": 0.4997,
      "num_input_tokens_seen": 5697776,
      "step": 8725
    },
    {
      "epoch": 4.5754716981132075,
      "grad_norm": 0.8111488223075867,
      "learning_rate": 4.7517555333539834e-05,
      "loss": 0.7551,
      "num_input_tokens_seen": 5700240,
      "step": 8730
    },
    {
      "epoch": 4.578092243186583,
      "grad_norm": 0.49385595321655273,
      "learning_rate": 4.751258550185284e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 5702992,
      "step": 8735
    },
    {
      "epoch": 4.580712788259958,
      "grad_norm": 0.9597947001457214,
      "learning_rate": 4.7507610960789804e-05,
      "loss": 0.4341,
      "num_input_tokens_seen": 5706544,
      "step": 8740
    },
    {
      "epoch": 4.583333333333333,
      "grad_norm": 0.531252920627594,
      "learning_rate": 4.7502631711391334e-05,
      "loss": 0.4751,
      "num_input_tokens_seen": 5709328,
      "step": 8745
    },
    {
      "epoch": 4.585953878406709,
      "grad_norm": 0.2890542447566986,
      "learning_rate": 4.749764775469904e-05,
      "loss": 0.4065,
      "num_input_tokens_seen": 5712720,
      "step": 8750
    },
    {
      "epoch": 4.588574423480084,
      "grad_norm": 0.4418117105960846,
      "learning_rate": 4.749265909175549e-05,
      "loss": 0.5831,
      "num_input_tokens_seen": 5715536,
      "step": 8755
    },
    {
      "epoch": 4.591194968553459,
      "grad_norm": 0.5393993854522705,
      "learning_rate": 4.7487665723604275e-05,
      "loss": 0.5026,
      "num_input_tokens_seen": 5718352,
      "step": 8760
    },
    {
      "epoch": 4.593815513626835,
      "grad_norm": 0.6060150861740112,
      "learning_rate": 4.7482667651289944e-05,
      "loss": 0.473,
      "num_input_tokens_seen": 5722096,
      "step": 8765
    },
    {
      "epoch": 4.59643605870021,
      "grad_norm": 0.8375265002250671,
      "learning_rate": 4.747766487585803e-05,
      "loss": 0.4604,
      "num_input_tokens_seen": 5724816,
      "step": 8770
    },
    {
      "epoch": 4.599056603773585,
      "grad_norm": 0.3511262834072113,
      "learning_rate": 4.747265739835507e-05,
      "loss": 0.4734,
      "num_input_tokens_seen": 5729040,
      "step": 8775
    },
    {
      "epoch": 4.60167714884696,
      "grad_norm": 0.37238025665283203,
      "learning_rate": 4.746764521982856e-05,
      "loss": 0.4546,
      "num_input_tokens_seen": 5732464,
      "step": 8780
    },
    {
      "epoch": 4.604297693920335,
      "grad_norm": 0.3876558840274811,
      "learning_rate": 4.7462628341326995e-05,
      "loss": 0.3381,
      "num_input_tokens_seen": 5735472,
      "step": 8785
    },
    {
      "epoch": 4.6069182389937104,
      "grad_norm": 0.3680097162723541,
      "learning_rate": 4.7457606763899856e-05,
      "loss": 0.4678,
      "num_input_tokens_seen": 5738896,
      "step": 8790
    },
    {
      "epoch": 4.609538784067086,
      "grad_norm": 0.6452273726463318,
      "learning_rate": 4.745258048859758e-05,
      "loss": 0.3931,
      "num_input_tokens_seen": 5742032,
      "step": 8795
    },
    {
      "epoch": 4.612159329140461,
      "grad_norm": 0.34110602736473083,
      "learning_rate": 4.744754951647162e-05,
      "loss": 0.5031,
      "num_input_tokens_seen": 5745168,
      "step": 8800
    },
    {
      "epoch": 4.614779874213837,
      "grad_norm": 0.46521106362342834,
      "learning_rate": 4.74425138485744e-05,
      "loss": 0.5407,
      "num_input_tokens_seen": 5747920,
      "step": 8805
    },
    {
      "epoch": 4.617400419287212,
      "grad_norm": 0.5103479623794556,
      "learning_rate": 4.743747348595932e-05,
      "loss": 0.4938,
      "num_input_tokens_seen": 5751120,
      "step": 8810
    },
    {
      "epoch": 4.620020964360587,
      "grad_norm": 0.4775232970714569,
      "learning_rate": 4.743242842968077e-05,
      "loss": 0.4513,
      "num_input_tokens_seen": 5754352,
      "step": 8815
    },
    {
      "epoch": 4.622641509433962,
      "grad_norm": 2.0764124393463135,
      "learning_rate": 4.742737868079411e-05,
      "loss": 0.4534,
      "num_input_tokens_seen": 5757200,
      "step": 8820
    },
    {
      "epoch": 4.6252620545073375,
      "grad_norm": 0.5018410086631775,
      "learning_rate": 4.7422324240355696e-05,
      "loss": 0.3588,
      "num_input_tokens_seen": 5760944,
      "step": 8825
    },
    {
      "epoch": 4.627882599580713,
      "grad_norm": 0.6111652851104736,
      "learning_rate": 4.741726510942286e-05,
      "loss": 0.5068,
      "num_input_tokens_seen": 5764848,
      "step": 8830
    },
    {
      "epoch": 4.630503144654088,
      "grad_norm": 0.49972283840179443,
      "learning_rate": 4.741220128905391e-05,
      "loss": 0.3996,
      "num_input_tokens_seen": 5767984,
      "step": 8835
    },
    {
      "epoch": 4.633123689727463,
      "grad_norm": 0.4845311939716339,
      "learning_rate": 4.740713278030815e-05,
      "loss": 0.4264,
      "num_input_tokens_seen": 5771472,
      "step": 8840
    },
    {
      "epoch": 4.635744234800838,
      "grad_norm": 0.748955488204956,
      "learning_rate": 4.740205958424584e-05,
      "loss": 0.5345,
      "num_input_tokens_seen": 5775536,
      "step": 8845
    },
    {
      "epoch": 4.638364779874214,
      "grad_norm": 0.34445956349372864,
      "learning_rate": 4.739698170192824e-05,
      "loss": 0.434,
      "num_input_tokens_seen": 5778704,
      "step": 8850
    },
    {
      "epoch": 4.640985324947589,
      "grad_norm": 0.5924513339996338,
      "learning_rate": 4.7391899134417596e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 5781104,
      "step": 8855
    },
    {
      "epoch": 4.643605870020965,
      "grad_norm": 0.5124208331108093,
      "learning_rate": 4.73868118827771e-05,
      "loss": 0.5007,
      "num_input_tokens_seen": 5784528,
      "step": 8860
    },
    {
      "epoch": 4.64622641509434,
      "grad_norm": 0.45085787773132324,
      "learning_rate": 4.738171994807097e-05,
      "loss": 0.7307,
      "num_input_tokens_seen": 5787184,
      "step": 8865
    },
    {
      "epoch": 4.648846960167715,
      "grad_norm": 0.8613166809082031,
      "learning_rate": 4.7376623331364364e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 5790032,
      "step": 8870
    },
    {
      "epoch": 4.65146750524109,
      "grad_norm": 0.4481440484523773,
      "learning_rate": 4.737152203372345e-05,
      "loss": 0.48,
      "num_input_tokens_seen": 5793104,
      "step": 8875
    },
    {
      "epoch": 4.654088050314465,
      "grad_norm": 0.764593243598938,
      "learning_rate": 4.7366416056215345e-05,
      "loss": 0.5668,
      "num_input_tokens_seen": 5796368,
      "step": 8880
    },
    {
      "epoch": 4.65670859538784,
      "grad_norm": 0.4995109736919403,
      "learning_rate": 4.7361305399908185e-05,
      "loss": 0.6233,
      "num_input_tokens_seen": 5799248,
      "step": 8885
    },
    {
      "epoch": 4.659329140461216,
      "grad_norm": 0.357639878988266,
      "learning_rate": 4.735619006587103e-05,
      "loss": 0.5523,
      "num_input_tokens_seen": 5802448,
      "step": 8890
    },
    {
      "epoch": 4.661949685534591,
      "grad_norm": 0.38148677349090576,
      "learning_rate": 4.735107005517398e-05,
      "loss": 0.511,
      "num_input_tokens_seen": 5805936,
      "step": 8895
    },
    {
      "epoch": 4.664570230607967,
      "grad_norm": 0.6858163475990295,
      "learning_rate": 4.734594536888806e-05,
      "loss": 0.4262,
      "num_input_tokens_seen": 5809744,
      "step": 8900
    },
    {
      "epoch": 4.667190775681342,
      "grad_norm": 0.4828437268733978,
      "learning_rate": 4.734081600808531e-05,
      "loss": 0.5739,
      "num_input_tokens_seen": 5812656,
      "step": 8905
    },
    {
      "epoch": 4.669811320754717,
      "grad_norm": 0.7362319231033325,
      "learning_rate": 4.733568197383872e-05,
      "loss": 0.6099,
      "num_input_tokens_seen": 5816048,
      "step": 8910
    },
    {
      "epoch": 4.672431865828092,
      "grad_norm": 0.29297879338264465,
      "learning_rate": 4.7330543267222294e-05,
      "loss": 0.5827,
      "num_input_tokens_seen": 5820080,
      "step": 8915
    },
    {
      "epoch": 4.6750524109014675,
      "grad_norm": 0.35808512568473816,
      "learning_rate": 4.732539988931097e-05,
      "loss": 0.518,
      "num_input_tokens_seen": 5823344,
      "step": 8920
    },
    {
      "epoch": 4.677672955974843,
      "grad_norm": 0.40421149134635925,
      "learning_rate": 4.732025184118068e-05,
      "loss": 0.4824,
      "num_input_tokens_seen": 5826448,
      "step": 8925
    },
    {
      "epoch": 4.680293501048218,
      "grad_norm": 0.739383339881897,
      "learning_rate": 4.731509912390837e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 5830000,
      "step": 8930
    },
    {
      "epoch": 4.682914046121593,
      "grad_norm": 0.32810714840888977,
      "learning_rate": 4.730994173857189e-05,
      "loss": 0.3589,
      "num_input_tokens_seen": 5832464,
      "step": 8935
    },
    {
      "epoch": 4.685534591194968,
      "grad_norm": 0.5081629157066345,
      "learning_rate": 4.7304779686250126e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 5835568,
      "step": 8940
    },
    {
      "epoch": 4.688155136268344,
      "grad_norm": 0.3250408470630646,
      "learning_rate": 4.729961296802292e-05,
      "loss": 0.4728,
      "num_input_tokens_seen": 5841360,
      "step": 8945
    },
    {
      "epoch": 4.690775681341719,
      "grad_norm": 0.5213723182678223,
      "learning_rate": 4.7294441584971095e-05,
      "loss": 0.3697,
      "num_input_tokens_seen": 5844240,
      "step": 8950
    },
    {
      "epoch": 4.693396226415095,
      "grad_norm": 0.4693278670310974,
      "learning_rate": 4.728926553817643e-05,
      "loss": 0.4143,
      "num_input_tokens_seen": 5848144,
      "step": 8955
    },
    {
      "epoch": 4.69601677148847,
      "grad_norm": 0.5811082124710083,
      "learning_rate": 4.728408482872171e-05,
      "loss": 0.5387,
      "num_input_tokens_seen": 5850800,
      "step": 8960
    },
    {
      "epoch": 4.698637316561845,
      "grad_norm": 0.39961984753608704,
      "learning_rate": 4.727889945769067e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 5853712,
      "step": 8965
    },
    {
      "epoch": 4.70125786163522,
      "grad_norm": 0.4459107518196106,
      "learning_rate": 4.727370942616803e-05,
      "loss": 0.4181,
      "num_input_tokens_seen": 5856400,
      "step": 8970
    },
    {
      "epoch": 4.703878406708595,
      "grad_norm": 0.6434398293495178,
      "learning_rate": 4.72685147352395e-05,
      "loss": 0.5203,
      "num_input_tokens_seen": 5859760,
      "step": 8975
    },
    {
      "epoch": 4.70649895178197,
      "grad_norm": 0.31291186809539795,
      "learning_rate": 4.7263315385991735e-05,
      "loss": 0.3891,
      "num_input_tokens_seen": 5862672,
      "step": 8980
    },
    {
      "epoch": 4.709119496855346,
      "grad_norm": 0.45200785994529724,
      "learning_rate": 4.7258111379512384e-05,
      "loss": 0.4988,
      "num_input_tokens_seen": 5866288,
      "step": 8985
    },
    {
      "epoch": 4.711740041928721,
      "grad_norm": 0.3983451724052429,
      "learning_rate": 4.725290271689007e-05,
      "loss": 0.5521,
      "num_input_tokens_seen": 5869424,
      "step": 8990
    },
    {
      "epoch": 4.714360587002097,
      "grad_norm": 0.2372683733701706,
      "learning_rate": 4.724768939921438e-05,
      "loss": 0.3607,
      "num_input_tokens_seen": 5872240,
      "step": 8995
    },
    {
      "epoch": 4.716981132075472,
      "grad_norm": 0.5104719996452332,
      "learning_rate": 4.7242471427575884e-05,
      "loss": 0.3925,
      "num_input_tokens_seen": 5874672,
      "step": 9000
    },
    {
      "epoch": 4.719601677148847,
      "grad_norm": 0.36906298995018005,
      "learning_rate": 4.723724880306612e-05,
      "loss": 0.4185,
      "num_input_tokens_seen": 5878544,
      "step": 9005
    },
    {
      "epoch": 4.722222222222222,
      "grad_norm": 0.41299378871917725,
      "learning_rate": 4.72320215267776e-05,
      "loss": 0.5302,
      "num_input_tokens_seen": 5882096,
      "step": 9010
    },
    {
      "epoch": 4.7248427672955975,
      "grad_norm": 0.7254588603973389,
      "learning_rate": 4.722678959980381e-05,
      "loss": 0.4958,
      "num_input_tokens_seen": 5886512,
      "step": 9015
    },
    {
      "epoch": 4.727463312368973,
      "grad_norm": 0.4030344486236572,
      "learning_rate": 4.7221553023239216e-05,
      "loss": 0.4218,
      "num_input_tokens_seen": 5889648,
      "step": 9020
    },
    {
      "epoch": 4.730083857442348,
      "grad_norm": 0.5768265724182129,
      "learning_rate": 4.721631179817925e-05,
      "loss": 0.469,
      "num_input_tokens_seen": 5892752,
      "step": 9025
    },
    {
      "epoch": 4.732704402515723,
      "grad_norm": 0.2589699625968933,
      "learning_rate": 4.721106592572031e-05,
      "loss": 0.6119,
      "num_input_tokens_seen": 5896112,
      "step": 9030
    },
    {
      "epoch": 4.735324947589098,
      "grad_norm": 0.3863157033920288,
      "learning_rate": 4.7205815406959774e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 5899440,
      "step": 9035
    },
    {
      "epoch": 4.737945492662474,
      "grad_norm": 0.5153785347938538,
      "learning_rate": 4.7200560242996e-05,
      "loss": 0.5755,
      "num_input_tokens_seen": 5902096,
      "step": 9040
    },
    {
      "epoch": 4.740566037735849,
      "grad_norm": 0.33460062742233276,
      "learning_rate": 4.719530043492829e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 5904880,
      "step": 9045
    },
    {
      "epoch": 4.743186582809225,
      "grad_norm": 0.5293139815330505,
      "learning_rate": 4.7190035983856954e-05,
      "loss": 0.505,
      "num_input_tokens_seen": 5907632,
      "step": 9050
    },
    {
      "epoch": 4.7458071278826,
      "grad_norm": 1.2158539295196533,
      "learning_rate": 4.718476689088325e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 5911120,
      "step": 9055
    },
    {
      "epoch": 4.748427672955975,
      "grad_norm": 0.4451080560684204,
      "learning_rate": 4.71794931571094e-05,
      "loss": 0.4239,
      "num_input_tokens_seen": 5914928,
      "step": 9060
    },
    {
      "epoch": 4.75104821802935,
      "grad_norm": 0.4406231939792633,
      "learning_rate": 4.717421478363862e-05,
      "loss": 0.3952,
      "num_input_tokens_seen": 5918192,
      "step": 9065
    },
    {
      "epoch": 4.753668763102725,
      "grad_norm": 0.37606561183929443,
      "learning_rate": 4.7168931771575085e-05,
      "loss": 0.3658,
      "num_input_tokens_seen": 5921232,
      "step": 9070
    },
    {
      "epoch": 4.7562893081761,
      "grad_norm": 0.28970837593078613,
      "learning_rate": 4.716364412202394e-05,
      "loss": 0.4435,
      "num_input_tokens_seen": 5925040,
      "step": 9075
    },
    {
      "epoch": 4.758909853249476,
      "grad_norm": 0.6742352843284607,
      "learning_rate": 4.7158351836091306e-05,
      "loss": 0.5016,
      "num_input_tokens_seen": 5927376,
      "step": 9080
    },
    {
      "epoch": 4.761530398322851,
      "grad_norm": 0.2956310212612152,
      "learning_rate": 4.7153054914884266e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 5930576,
      "step": 9085
    },
    {
      "epoch": 4.764150943396227,
      "grad_norm": 2.2580161094665527,
      "learning_rate": 4.714775335951086e-05,
      "loss": 0.3877,
      "num_input_tokens_seen": 5933328,
      "step": 9090
    },
    {
      "epoch": 4.766771488469602,
      "grad_norm": 0.36696916818618774,
      "learning_rate": 4.7142447171080137e-05,
      "loss": 0.4757,
      "num_input_tokens_seen": 5936208,
      "step": 9095
    },
    {
      "epoch": 4.769392033542977,
      "grad_norm": 0.4246678352355957,
      "learning_rate": 4.713713635070208e-05,
      "loss": 0.6166,
      "num_input_tokens_seen": 5938832,
      "step": 9100
    },
    {
      "epoch": 4.772012578616352,
      "grad_norm": 0.6299715042114258,
      "learning_rate": 4.713182089948765e-05,
      "loss": 0.3826,
      "num_input_tokens_seen": 5942128,
      "step": 9105
    },
    {
      "epoch": 4.7746331236897275,
      "grad_norm": 0.3337019681930542,
      "learning_rate": 4.712650081854878e-05,
      "loss": 0.2992,
      "num_input_tokens_seen": 5945648,
      "step": 9110
    },
    {
      "epoch": 4.777253668763103,
      "grad_norm": 0.2802993357181549,
      "learning_rate": 4.712117610899837e-05,
      "loss": 0.4025,
      "num_input_tokens_seen": 5950352,
      "step": 9115
    },
    {
      "epoch": 4.779874213836478,
      "grad_norm": 0.6369849443435669,
      "learning_rate": 4.7115846771950286e-05,
      "loss": 0.5974,
      "num_input_tokens_seen": 5953488,
      "step": 9120
    },
    {
      "epoch": 4.782494758909853,
      "grad_norm": 0.3914448618888855,
      "learning_rate": 4.711051280851938e-05,
      "loss": 0.449,
      "num_input_tokens_seen": 5956784,
      "step": 9125
    },
    {
      "epoch": 4.785115303983228,
      "grad_norm": 0.49069663882255554,
      "learning_rate": 4.7105174219821435e-05,
      "loss": 0.5196,
      "num_input_tokens_seen": 5960080,
      "step": 9130
    },
    {
      "epoch": 4.787735849056604,
      "grad_norm": 0.34800949692726135,
      "learning_rate": 4.709983100697323e-05,
      "loss": 0.3751,
      "num_input_tokens_seen": 5962864,
      "step": 9135
    },
    {
      "epoch": 4.790356394129979,
      "grad_norm": 0.33314236998558044,
      "learning_rate": 4.709448317109252e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 5966000,
      "step": 9140
    },
    {
      "epoch": 4.7929769392033545,
      "grad_norm": 0.4042744040489197,
      "learning_rate": 4.708913071329798e-05,
      "loss": 0.5219,
      "num_input_tokens_seen": 5969168,
      "step": 9145
    },
    {
      "epoch": 4.79559748427673,
      "grad_norm": 0.31151869893074036,
      "learning_rate": 4.708377363470932e-05,
      "loss": 0.5085,
      "num_input_tokens_seen": 5972720,
      "step": 9150
    },
    {
      "epoch": 4.798218029350105,
      "grad_norm": 0.2788299024105072,
      "learning_rate": 4.7078411936447144e-05,
      "loss": 0.3554,
      "num_input_tokens_seen": 5975760,
      "step": 9155
    },
    {
      "epoch": 4.80083857442348,
      "grad_norm": 0.4787103831768036,
      "learning_rate": 4.707304561963308e-05,
      "loss": 0.3575,
      "num_input_tokens_seen": 5979120,
      "step": 9160
    },
    {
      "epoch": 4.803459119496855,
      "grad_norm": 0.40041854977607727,
      "learning_rate": 4.70676746853897e-05,
      "loss": 0.5054,
      "num_input_tokens_seen": 5983376,
      "step": 9165
    },
    {
      "epoch": 4.80607966457023,
      "grad_norm": 0.43813198804855347,
      "learning_rate": 4.706229913484054e-05,
      "loss": 0.5579,
      "num_input_tokens_seen": 5986416,
      "step": 9170
    },
    {
      "epoch": 4.808700209643606,
      "grad_norm": 0.5404593348503113,
      "learning_rate": 4.705691896911009e-05,
      "loss": 0.3939,
      "num_input_tokens_seen": 5989136,
      "step": 9175
    },
    {
      "epoch": 4.811320754716981,
      "grad_norm": 0.7221613526344299,
      "learning_rate": 4.705153418932383e-05,
      "loss": 0.6572,
      "num_input_tokens_seen": 5992112,
      "step": 9180
    },
    {
      "epoch": 4.813941299790356,
      "grad_norm": 0.3652460277080536,
      "learning_rate": 4.7046144796608206e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 5995472,
      "step": 9185
    },
    {
      "epoch": 4.816561844863732,
      "grad_norm": 0.5058136582374573,
      "learning_rate": 4.70407507920906e-05,
      "loss": 0.4353,
      "num_input_tokens_seen": 5998736,
      "step": 9190
    },
    {
      "epoch": 4.819182389937107,
      "grad_norm": 0.4173288345336914,
      "learning_rate": 4.7035352176899384e-05,
      "loss": 0.6604,
      "num_input_tokens_seen": 6001552,
      "step": 9195
    },
    {
      "epoch": 4.821802935010482,
      "grad_norm": 0.8758785128593445,
      "learning_rate": 4.7029948952163885e-05,
      "loss": 0.5189,
      "num_input_tokens_seen": 6005520,
      "step": 9200
    },
    {
      "epoch": 4.8244234800838575,
      "grad_norm": 0.3321840465068817,
      "learning_rate": 4.7024541119014386e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 6009200,
      "step": 9205
    },
    {
      "epoch": 4.827044025157233,
      "grad_norm": 0.36759573221206665,
      "learning_rate": 4.701912867858217e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 6012912,
      "step": 9210
    },
    {
      "epoch": 4.829664570230608,
      "grad_norm": 0.35079771280288696,
      "learning_rate": 4.7013711631999434e-05,
      "loss": 0.5582,
      "num_input_tokens_seen": 6017744,
      "step": 9215
    },
    {
      "epoch": 4.832285115303983,
      "grad_norm": 0.8061214089393616,
      "learning_rate": 4.700828998039937e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 6020624,
      "step": 9220
    },
    {
      "epoch": 4.834905660377358,
      "grad_norm": 0.3080458641052246,
      "learning_rate": 4.700286372491612e-05,
      "loss": 0.4412,
      "num_input_tokens_seen": 6023312,
      "step": 9225
    },
    {
      "epoch": 4.837526205450734,
      "grad_norm": 0.979871928691864,
      "learning_rate": 4.6997432866684816e-05,
      "loss": 0.4782,
      "num_input_tokens_seen": 6026032,
      "step": 9230
    },
    {
      "epoch": 4.840146750524109,
      "grad_norm": 0.5835301280021667,
      "learning_rate": 4.6991997406841504e-05,
      "loss": 0.4012,
      "num_input_tokens_seen": 6029872,
      "step": 9235
    },
    {
      "epoch": 4.8427672955974845,
      "grad_norm": 0.2027595192193985,
      "learning_rate": 4.698655734652324e-05,
      "loss": 0.4712,
      "num_input_tokens_seen": 6033232,
      "step": 9240
    },
    {
      "epoch": 4.84538784067086,
      "grad_norm": 1.5700222253799438,
      "learning_rate": 4.698111268686801e-05,
      "loss": 0.6568,
      "num_input_tokens_seen": 6035888,
      "step": 9245
    },
    {
      "epoch": 4.848008385744235,
      "grad_norm": 0.28617820143699646,
      "learning_rate": 4.6975663429014785e-05,
      "loss": 0.4636,
      "num_input_tokens_seen": 6039408,
      "step": 9250
    },
    {
      "epoch": 4.85062893081761,
      "grad_norm": 0.35618942975997925,
      "learning_rate": 4.697020957410347e-05,
      "loss": 0.5444,
      "num_input_tokens_seen": 6042736,
      "step": 9255
    },
    {
      "epoch": 4.853249475890985,
      "grad_norm": 0.49742090702056885,
      "learning_rate": 4.6964751123274974e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 6045296,
      "step": 9260
    },
    {
      "epoch": 4.85587002096436,
      "grad_norm": 0.5499246716499329,
      "learning_rate": 4.695928807767113e-05,
      "loss": 0.4766,
      "num_input_tokens_seen": 6047920,
      "step": 9265
    },
    {
      "epoch": 4.8584905660377355,
      "grad_norm": 0.29366061091423035,
      "learning_rate": 4.695382043843474e-05,
      "loss": 0.3681,
      "num_input_tokens_seen": 6052432,
      "step": 9270
    },
    {
      "epoch": 4.861111111111111,
      "grad_norm": 0.6978223323822021,
      "learning_rate": 4.694834820670958e-05,
      "loss": 0.4407,
      "num_input_tokens_seen": 6055472,
      "step": 9275
    },
    {
      "epoch": 4.863731656184486,
      "grad_norm": 0.5401312112808228,
      "learning_rate": 4.694287138364037e-05,
      "loss": 0.5831,
      "num_input_tokens_seen": 6059184,
      "step": 9280
    },
    {
      "epoch": 4.866352201257862,
      "grad_norm": 0.5552365779876709,
      "learning_rate": 4.6937389970372815e-05,
      "loss": 0.5999,
      "num_input_tokens_seen": 6062192,
      "step": 9285
    },
    {
      "epoch": 4.868972746331237,
      "grad_norm": 0.5835473537445068,
      "learning_rate": 4.693190396805354e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 6065264,
      "step": 9290
    },
    {
      "epoch": 4.871593291404612,
      "grad_norm": 0.35950085520744324,
      "learning_rate": 4.692641337783017e-05,
      "loss": 0.5353,
      "num_input_tokens_seen": 6069008,
      "step": 9295
    },
    {
      "epoch": 4.8742138364779874,
      "grad_norm": 0.3165225088596344,
      "learning_rate": 4.6920918200851274e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 6072624,
      "step": 9300
    },
    {
      "epoch": 4.876834381551363,
      "grad_norm": 0.4007350504398346,
      "learning_rate": 4.691541843826638e-05,
      "loss": 0.584,
      "num_input_tokens_seen": 6075280,
      "step": 9305
    },
    {
      "epoch": 4.879454926624738,
      "grad_norm": 0.7930954694747925,
      "learning_rate": 4.690991409122597e-05,
      "loss": 0.4861,
      "num_input_tokens_seen": 6078448,
      "step": 9310
    },
    {
      "epoch": 4.882075471698113,
      "grad_norm": 0.664094090461731,
      "learning_rate": 4.690440516088149e-05,
      "loss": 0.3094,
      "num_input_tokens_seen": 6081008,
      "step": 9315
    },
    {
      "epoch": 4.884696016771488,
      "grad_norm": 0.342613160610199,
      "learning_rate": 4.689889164838536e-05,
      "loss": 0.4861,
      "num_input_tokens_seen": 6085296,
      "step": 9320
    },
    {
      "epoch": 4.887316561844864,
      "grad_norm": 0.44652169942855835,
      "learning_rate": 4.6893373554890925e-05,
      "loss": 0.4799,
      "num_input_tokens_seen": 6087664,
      "step": 9325
    },
    {
      "epoch": 4.889937106918239,
      "grad_norm": 0.5268837809562683,
      "learning_rate": 4.688785088155251e-05,
      "loss": 0.4741,
      "num_input_tokens_seen": 6090832,
      "step": 9330
    },
    {
      "epoch": 4.8925576519916145,
      "grad_norm": 0.6258094310760498,
      "learning_rate": 4.688232362952541e-05,
      "loss": 0.533,
      "num_input_tokens_seen": 6093616,
      "step": 9335
    },
    {
      "epoch": 4.89517819706499,
      "grad_norm": 0.44531649351119995,
      "learning_rate": 4.687679179996585e-05,
      "loss": 0.3723,
      "num_input_tokens_seen": 6096592,
      "step": 9340
    },
    {
      "epoch": 4.897798742138365,
      "grad_norm": 0.232569620013237,
      "learning_rate": 4.687125539403103e-05,
      "loss": 0.3324,
      "num_input_tokens_seen": 6100144,
      "step": 9345
    },
    {
      "epoch": 4.90041928721174,
      "grad_norm": 0.5516849756240845,
      "learning_rate": 4.68657144128791e-05,
      "loss": 0.5284,
      "num_input_tokens_seen": 6103536,
      "step": 9350
    },
    {
      "epoch": 4.903039832285115,
      "grad_norm": 0.30140426754951477,
      "learning_rate": 4.6860168857669175e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 6107408,
      "step": 9355
    },
    {
      "epoch": 4.90566037735849,
      "grad_norm": 0.40345045924186707,
      "learning_rate": 4.6854618729561315e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 6110480,
      "step": 9360
    },
    {
      "epoch": 4.9082809224318655,
      "grad_norm": 0.682110071182251,
      "learning_rate": 4.684906402971655e-05,
      "loss": 0.4063,
      "num_input_tokens_seen": 6113488,
      "step": 9365
    },
    {
      "epoch": 4.910901467505241,
      "grad_norm": 0.3985036611557007,
      "learning_rate": 4.684350475929685e-05,
      "loss": 0.4753,
      "num_input_tokens_seen": 6117648,
      "step": 9370
    },
    {
      "epoch": 4.913522012578616,
      "grad_norm": 0.7254964113235474,
      "learning_rate": 4.683794091946517e-05,
      "loss": 0.4672,
      "num_input_tokens_seen": 6120592,
      "step": 9375
    },
    {
      "epoch": 4.916142557651992,
      "grad_norm": 0.5441609621047974,
      "learning_rate": 4.683237251138538e-05,
      "loss": 0.5264,
      "num_input_tokens_seen": 6124400,
      "step": 9380
    },
    {
      "epoch": 4.918763102725367,
      "grad_norm": 0.3994390070438385,
      "learning_rate": 4.6826799536222346e-05,
      "loss": 0.5092,
      "num_input_tokens_seen": 6128400,
      "step": 9385
    },
    {
      "epoch": 4.921383647798742,
      "grad_norm": 0.3122919201850891,
      "learning_rate": 4.6821221995141865e-05,
      "loss": 0.4607,
      "num_input_tokens_seen": 6131536,
      "step": 9390
    },
    {
      "epoch": 4.924004192872117,
      "grad_norm": 0.3265700936317444,
      "learning_rate": 4.681563988931068e-05,
      "loss": 0.33,
      "num_input_tokens_seen": 6135280,
      "step": 9395
    },
    {
      "epoch": 4.926624737945493,
      "grad_norm": 0.7722629308700562,
      "learning_rate": 4.681005321989652e-05,
      "loss": 0.5051,
      "num_input_tokens_seen": 6137680,
      "step": 9400
    },
    {
      "epoch": 4.929245283018868,
      "grad_norm": 0.5320643782615662,
      "learning_rate": 4.6804461988068055e-05,
      "loss": 0.6975,
      "num_input_tokens_seen": 6141264,
      "step": 9405
    },
    {
      "epoch": 4.931865828092243,
      "grad_norm": 0.759383499622345,
      "learning_rate": 4.6798866194994896e-05,
      "loss": 0.4192,
      "num_input_tokens_seen": 6145104,
      "step": 9410
    },
    {
      "epoch": 4.934486373165618,
      "grad_norm": 0.3136334717273712,
      "learning_rate": 4.6793265841847624e-05,
      "loss": 0.453,
      "num_input_tokens_seen": 6148976,
      "step": 9415
    },
    {
      "epoch": 4.937106918238994,
      "grad_norm": 0.5818079710006714,
      "learning_rate": 4.6787660929797766e-05,
      "loss": 0.6299,
      "num_input_tokens_seen": 6152560,
      "step": 9420
    },
    {
      "epoch": 4.939727463312369,
      "grad_norm": 0.5892509818077087,
      "learning_rate": 4.678205146001781e-05,
      "loss": 0.5729,
      "num_input_tokens_seen": 6155024,
      "step": 9425
    },
    {
      "epoch": 4.9423480083857445,
      "grad_norm": 0.3734723925590515,
      "learning_rate": 4.67764374336812e-05,
      "loss": 0.5,
      "num_input_tokens_seen": 6157584,
      "step": 9430
    },
    {
      "epoch": 4.94496855345912,
      "grad_norm": 0.5727794766426086,
      "learning_rate": 4.6770818851962305e-05,
      "loss": 0.5148,
      "num_input_tokens_seen": 6160240,
      "step": 9435
    },
    {
      "epoch": 4.947589098532495,
      "grad_norm": 0.34131109714508057,
      "learning_rate": 4.676519571603648e-05,
      "loss": 0.5181,
      "num_input_tokens_seen": 6163696,
      "step": 9440
    },
    {
      "epoch": 4.95020964360587,
      "grad_norm": 0.3098919093608856,
      "learning_rate": 4.6759568027080034e-05,
      "loss": 0.3601,
      "num_input_tokens_seen": 6168880,
      "step": 9445
    },
    {
      "epoch": 4.952830188679245,
      "grad_norm": 0.6158428192138672,
      "learning_rate": 4.67539357862702e-05,
      "loss": 0.4953,
      "num_input_tokens_seen": 6172208,
      "step": 9450
    },
    {
      "epoch": 4.95545073375262,
      "grad_norm": 0.37639811635017395,
      "learning_rate": 4.6748298994785175e-05,
      "loss": 0.3611,
      "num_input_tokens_seen": 6174928,
      "step": 9455
    },
    {
      "epoch": 4.9580712788259955,
      "grad_norm": 0.30324864387512207,
      "learning_rate": 4.6742657653804116e-05,
      "loss": 0.4708,
      "num_input_tokens_seen": 6178992,
      "step": 9460
    },
    {
      "epoch": 4.960691823899371,
      "grad_norm": 0.4857468605041504,
      "learning_rate": 4.673701176450713e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 6182128,
      "step": 9465
    },
    {
      "epoch": 4.963312368972746,
      "grad_norm": 0.6579537987709045,
      "learning_rate": 4.6731361328075286e-05,
      "loss": 0.4343,
      "num_input_tokens_seen": 6185328,
      "step": 9470
    },
    {
      "epoch": 4.965932914046122,
      "grad_norm": 0.2971363365650177,
      "learning_rate": 4.672570634569056e-05,
      "loss": 0.4769,
      "num_input_tokens_seen": 6188560,
      "step": 9475
    },
    {
      "epoch": 4.968553459119497,
      "grad_norm": 0.5022187232971191,
      "learning_rate": 4.672004681853593e-05,
      "loss": 0.56,
      "num_input_tokens_seen": 6193136,
      "step": 9480
    },
    {
      "epoch": 4.971174004192872,
      "grad_norm": 0.4006801247596741,
      "learning_rate": 4.6714382747795304e-05,
      "loss": 0.6804,
      "num_input_tokens_seen": 6196816,
      "step": 9485
    },
    {
      "epoch": 4.973794549266247,
      "grad_norm": 0.5377445816993713,
      "learning_rate": 4.670871413465354e-05,
      "loss": 0.4504,
      "num_input_tokens_seen": 6200272,
      "step": 9490
    },
    {
      "epoch": 4.976415094339623,
      "grad_norm": 1.0111337900161743,
      "learning_rate": 4.670304098029645e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 6203216,
      "step": 9495
    },
    {
      "epoch": 4.979035639412998,
      "grad_norm": 0.33985447883605957,
      "learning_rate": 4.669736328591078e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 6206736,
      "step": 9500
    },
    {
      "epoch": 4.981656184486373,
      "grad_norm": 0.2906463146209717,
      "learning_rate": 4.669168105268424e-05,
      "loss": 0.4034,
      "num_input_tokens_seen": 6209936,
      "step": 9505
    },
    {
      "epoch": 4.984276729559748,
      "grad_norm": 0.3168613016605377,
      "learning_rate": 4.6685994281805514e-05,
      "loss": 0.4791,
      "num_input_tokens_seen": 6213424,
      "step": 9510
    },
    {
      "epoch": 4.986897274633124,
      "grad_norm": 0.6818738579750061,
      "learning_rate": 4.6680302974464175e-05,
      "loss": 0.3946,
      "num_input_tokens_seen": 6217360,
      "step": 9515
    },
    {
      "epoch": 4.989517819706499,
      "grad_norm": 0.6198022961616516,
      "learning_rate": 4.66746071318508e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 6219984,
      "step": 9520
    },
    {
      "epoch": 4.9921383647798745,
      "grad_norm": 0.2935265004634857,
      "learning_rate": 4.6668906755156896e-05,
      "loss": 0.3959,
      "num_input_tokens_seen": 6223664,
      "step": 9525
    },
    {
      "epoch": 4.99475890985325,
      "grad_norm": 0.3634566366672516,
      "learning_rate": 4.6663201845574906e-05,
      "loss": 0.4832,
      "num_input_tokens_seen": 6227440,
      "step": 9530
    },
    {
      "epoch": 4.997379454926625,
      "grad_norm": 0.46459269523620605,
      "learning_rate": 4.6657492404298234e-05,
      "loss": 0.4064,
      "num_input_tokens_seen": 6230608,
      "step": 9535
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.7430330514907837,
      "learning_rate": 4.665177843252124e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 6233920,
      "step": 9540
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.487842857837677,
      "eval_runtime": 9.7016,
      "eval_samples_per_second": 87.408,
      "eval_steps_per_second": 21.852,
      "num_input_tokens_seen": 6233920,
      "step": 9540
    },
    {
      "epoch": 5.002620545073375,
      "grad_norm": 0.40880003571510315,
      "learning_rate": 4.664605993143921e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 6237312,
      "step": 9545
    },
    {
      "epoch": 5.00524109014675,
      "grad_norm": 0.4319440424442291,
      "learning_rate": 4.6640336902248396e-05,
      "loss": 0.4084,
      "num_input_tokens_seen": 6241344,
      "step": 9550
    },
    {
      "epoch": 5.0078616352201255,
      "grad_norm": 0.5077528953552246,
      "learning_rate": 4.6634609346145984e-05,
      "loss": 0.5562,
      "num_input_tokens_seen": 6244064,
      "step": 9555
    },
    {
      "epoch": 5.010482180293501,
      "grad_norm": 0.3759840428829193,
      "learning_rate": 4.6628877264330116e-05,
      "loss": 0.6503,
      "num_input_tokens_seen": 6247008,
      "step": 9560
    },
    {
      "epoch": 5.013102725366877,
      "grad_norm": 0.402874618768692,
      "learning_rate": 4.6623140657999886e-05,
      "loss": 0.5659,
      "num_input_tokens_seen": 6250016,
      "step": 9565
    },
    {
      "epoch": 5.015723270440252,
      "grad_norm": 0.3026523292064667,
      "learning_rate": 4.661739952835532e-05,
      "loss": 0.5874,
      "num_input_tokens_seen": 6253504,
      "step": 9570
    },
    {
      "epoch": 5.018343815513627,
      "grad_norm": 0.48932644724845886,
      "learning_rate": 4.66116538765974e-05,
      "loss": 0.4892,
      "num_input_tokens_seen": 6256256,
      "step": 9575
    },
    {
      "epoch": 5.020964360587002,
      "grad_norm": 0.34245073795318604,
      "learning_rate": 4.6605903703928034e-05,
      "loss": 0.639,
      "num_input_tokens_seen": 6260160,
      "step": 9580
    },
    {
      "epoch": 5.023584905660377,
      "grad_norm": 0.38653701543807983,
      "learning_rate": 4.6600149011550116e-05,
      "loss": 0.4579,
      "num_input_tokens_seen": 6264032,
      "step": 9585
    },
    {
      "epoch": 5.026205450733753,
      "grad_norm": 0.2990507483482361,
      "learning_rate": 4.659438980066745e-05,
      "loss": 0.5246,
      "num_input_tokens_seen": 6267552,
      "step": 9590
    },
    {
      "epoch": 5.028825995807128,
      "grad_norm": 0.3118433654308319,
      "learning_rate": 4.65886260724848e-05,
      "loss": 0.4843,
      "num_input_tokens_seen": 6270272,
      "step": 9595
    },
    {
      "epoch": 5.031446540880503,
      "grad_norm": 0.35911935567855835,
      "learning_rate": 4.658285782820787e-05,
      "loss": 0.2928,
      "num_input_tokens_seen": 6274624,
      "step": 9600
    },
    {
      "epoch": 5.034067085953878,
      "grad_norm": 0.45756205916404724,
      "learning_rate": 4.657708506904331e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 6277728,
      "step": 9605
    },
    {
      "epoch": 5.036687631027253,
      "grad_norm": 0.9460674524307251,
      "learning_rate": 4.6571307796198724e-05,
      "loss": 0.4473,
      "num_input_tokens_seen": 6281024,
      "step": 9610
    },
    {
      "epoch": 5.039308176100629,
      "grad_norm": 0.5045081377029419,
      "learning_rate": 4.656552601088265e-05,
      "loss": 0.588,
      "num_input_tokens_seen": 6284576,
      "step": 9615
    },
    {
      "epoch": 5.0419287211740045,
      "grad_norm": 0.4254765212535858,
      "learning_rate": 4.655973971430456e-05,
      "loss": 0.5524,
      "num_input_tokens_seen": 6288096,
      "step": 9620
    },
    {
      "epoch": 5.04454926624738,
      "grad_norm": 0.5158745646476746,
      "learning_rate": 4.6553948907674895e-05,
      "loss": 0.5236,
      "num_input_tokens_seen": 6292128,
      "step": 9625
    },
    {
      "epoch": 5.047169811320755,
      "grad_norm": 0.6685813665390015,
      "learning_rate": 4.654815359220501e-05,
      "loss": 0.4858,
      "num_input_tokens_seen": 6298912,
      "step": 9630
    },
    {
      "epoch": 5.04979035639413,
      "grad_norm": 0.4678911864757538,
      "learning_rate": 4.6542353769107226e-05,
      "loss": 0.4729,
      "num_input_tokens_seen": 6301536,
      "step": 9635
    },
    {
      "epoch": 5.052410901467505,
      "grad_norm": 0.42605695128440857,
      "learning_rate": 4.653654943959481e-05,
      "loss": 0.379,
      "num_input_tokens_seen": 6303936,
      "step": 9640
    },
    {
      "epoch": 5.05503144654088,
      "grad_norm": 0.631657600402832,
      "learning_rate": 4.653074060488195e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 6306880,
      "step": 9645
    },
    {
      "epoch": 5.0576519916142555,
      "grad_norm": 0.41566216945648193,
      "learning_rate": 4.6524927266183796e-05,
      "loss": 0.4796,
      "num_input_tokens_seen": 6309920,
      "step": 9650
    },
    {
      "epoch": 5.060272536687631,
      "grad_norm": 0.47156763076782227,
      "learning_rate": 4.6519109424716426e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 6312480,
      "step": 9655
    },
    {
      "epoch": 5.062893081761007,
      "grad_norm": 0.2973291575908661,
      "learning_rate": 4.6513287081696855e-05,
      "loss": 0.4308,
      "num_input_tokens_seen": 6316256,
      "step": 9660
    },
    {
      "epoch": 5.065513626834382,
      "grad_norm": 0.4579954445362091,
      "learning_rate": 4.6507460238343076e-05,
      "loss": 0.5225,
      "num_input_tokens_seen": 6319520,
      "step": 9665
    },
    {
      "epoch": 5.068134171907757,
      "grad_norm": 0.35063210129737854,
      "learning_rate": 4.6501628895873984e-05,
      "loss": 0.5734,
      "num_input_tokens_seen": 6323328,
      "step": 9670
    },
    {
      "epoch": 5.070754716981132,
      "grad_norm": 0.29103410243988037,
      "learning_rate": 4.649579305550942e-05,
      "loss": 0.3079,
      "num_input_tokens_seen": 6326720,
      "step": 9675
    },
    {
      "epoch": 5.073375262054507,
      "grad_norm": 0.7303856015205383,
      "learning_rate": 4.64899527184702e-05,
      "loss": 0.4768,
      "num_input_tokens_seen": 6329280,
      "step": 9680
    },
    {
      "epoch": 5.075995807127883,
      "grad_norm": 0.2970055937767029,
      "learning_rate": 4.6484107885978035e-05,
      "loss": 0.4278,
      "num_input_tokens_seen": 6332384,
      "step": 9685
    },
    {
      "epoch": 5.078616352201258,
      "grad_norm": 0.5211070775985718,
      "learning_rate": 4.64782585592556e-05,
      "loss": 0.3675,
      "num_input_tokens_seen": 6336096,
      "step": 9690
    },
    {
      "epoch": 5.081236897274633,
      "grad_norm": 0.6640990972518921,
      "learning_rate": 4.647240473952652e-05,
      "loss": 0.5655,
      "num_input_tokens_seen": 6338432,
      "step": 9695
    },
    {
      "epoch": 5.083857442348008,
      "grad_norm": 0.5580735206604004,
      "learning_rate": 4.6466546428015336e-05,
      "loss": 0.4634,
      "num_input_tokens_seen": 6342016,
      "step": 9700
    },
    {
      "epoch": 5.086477987421383,
      "grad_norm": 0.980336606502533,
      "learning_rate": 4.646068362594753e-05,
      "loss": 0.5472,
      "num_input_tokens_seen": 6344896,
      "step": 9705
    },
    {
      "epoch": 5.089098532494759,
      "grad_norm": 0.4001386761665344,
      "learning_rate": 4.645481633454956e-05,
      "loss": 0.4887,
      "num_input_tokens_seen": 6348096,
      "step": 9710
    },
    {
      "epoch": 5.0917190775681345,
      "grad_norm": 0.6710484027862549,
      "learning_rate": 4.644894455504878e-05,
      "loss": 0.5794,
      "num_input_tokens_seen": 6351488,
      "step": 9715
    },
    {
      "epoch": 5.09433962264151,
      "grad_norm": 0.4695870578289032,
      "learning_rate": 4.64430682886735e-05,
      "loss": 0.5434,
      "num_input_tokens_seen": 6354464,
      "step": 9720
    },
    {
      "epoch": 5.096960167714885,
      "grad_norm": 0.3636413514614105,
      "learning_rate": 4.643718753665297e-05,
      "loss": 0.454,
      "num_input_tokens_seen": 6358912,
      "step": 9725
    },
    {
      "epoch": 5.09958071278826,
      "grad_norm": 1.199541449546814,
      "learning_rate": 4.643130230021737e-05,
      "loss": 0.397,
      "num_input_tokens_seen": 6362112,
      "step": 9730
    },
    {
      "epoch": 5.102201257861635,
      "grad_norm": 0.4298607110977173,
      "learning_rate": 4.6425412580597836e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 6364896,
      "step": 9735
    },
    {
      "epoch": 5.10482180293501,
      "grad_norm": 0.4718182384967804,
      "learning_rate": 4.641951837902643e-05,
      "loss": 0.6578,
      "num_input_tokens_seen": 6368160,
      "step": 9740
    },
    {
      "epoch": 5.1074423480083855,
      "grad_norm": 0.4307939410209656,
      "learning_rate": 4.641361969673614e-05,
      "loss": 0.3807,
      "num_input_tokens_seen": 6371264,
      "step": 9745
    },
    {
      "epoch": 5.110062893081761,
      "grad_norm": 0.4345947206020355,
      "learning_rate": 4.640771653496091e-05,
      "loss": 0.53,
      "num_input_tokens_seen": 6374144,
      "step": 9750
    },
    {
      "epoch": 5.112683438155136,
      "grad_norm": 0.4280594289302826,
      "learning_rate": 4.64018088949356e-05,
      "loss": 0.4508,
      "num_input_tokens_seen": 6377600,
      "step": 9755
    },
    {
      "epoch": 5.115303983228512,
      "grad_norm": 0.7750841379165649,
      "learning_rate": 4.6395896777896055e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 6380608,
      "step": 9760
    },
    {
      "epoch": 5.117924528301887,
      "grad_norm": 0.43036147952079773,
      "learning_rate": 4.638998018507899e-05,
      "loss": 0.6094,
      "num_input_tokens_seen": 6384640,
      "step": 9765
    },
    {
      "epoch": 5.120545073375262,
      "grad_norm": 0.6729573011398315,
      "learning_rate": 4.63840591177221e-05,
      "loss": 0.4096,
      "num_input_tokens_seen": 6387456,
      "step": 9770
    },
    {
      "epoch": 5.123165618448637,
      "grad_norm": 0.42943233251571655,
      "learning_rate": 4.637813357706401e-05,
      "loss": 0.4525,
      "num_input_tokens_seen": 6390880,
      "step": 9775
    },
    {
      "epoch": 5.1257861635220126,
      "grad_norm": 0.46295487880706787,
      "learning_rate": 4.6372203564344264e-05,
      "loss": 0.3801,
      "num_input_tokens_seen": 6393696,
      "step": 9780
    },
    {
      "epoch": 5.128406708595388,
      "grad_norm": 0.2769376039505005,
      "learning_rate": 4.6366269080803364e-05,
      "loss": 0.4825,
      "num_input_tokens_seen": 6397472,
      "step": 9785
    },
    {
      "epoch": 5.131027253668763,
      "grad_norm": 0.3108634352684021,
      "learning_rate": 4.636033012768273e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 6400576,
      "step": 9790
    },
    {
      "epoch": 5.133647798742138,
      "grad_norm": 0.37657397985458374,
      "learning_rate": 4.635438670622473e-05,
      "loss": 0.5889,
      "num_input_tokens_seen": 6403168,
      "step": 9795
    },
    {
      "epoch": 5.136268343815513,
      "grad_norm": 0.3167400360107422,
      "learning_rate": 4.6348438817672646e-05,
      "loss": 0.4934,
      "num_input_tokens_seen": 6406144,
      "step": 9800
    },
    {
      "epoch": 5.138888888888889,
      "grad_norm": 0.5608848929405212,
      "learning_rate": 4.634248646327072e-05,
      "loss": 0.5533,
      "num_input_tokens_seen": 6409344,
      "step": 9805
    },
    {
      "epoch": 5.1415094339622645,
      "grad_norm": 0.6569456458091736,
      "learning_rate": 4.6336529644264114e-05,
      "loss": 0.45,
      "num_input_tokens_seen": 6411776,
      "step": 9810
    },
    {
      "epoch": 5.14412997903564,
      "grad_norm": 0.26909029483795166,
      "learning_rate": 4.633056836189893e-05,
      "loss": 0.3569,
      "num_input_tokens_seen": 6414784,
      "step": 9815
    },
    {
      "epoch": 5.146750524109015,
      "grad_norm": 1.0206314325332642,
      "learning_rate": 4.632460261742219e-05,
      "loss": 0.4138,
      "num_input_tokens_seen": 6417248,
      "step": 9820
    },
    {
      "epoch": 5.14937106918239,
      "grad_norm": 0.7421773672103882,
      "learning_rate": 4.631863241208187e-05,
      "loss": 0.586,
      "num_input_tokens_seen": 6420416,
      "step": 9825
    },
    {
      "epoch": 5.151991614255765,
      "grad_norm": 0.5902373194694519,
      "learning_rate": 4.6312657747126865e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 6424288,
      "step": 9830
    },
    {
      "epoch": 5.15461215932914,
      "grad_norm": 0.4623020589351654,
      "learning_rate": 4.630667862380701e-05,
      "loss": 0.566,
      "num_input_tokens_seen": 6426784,
      "step": 9835
    },
    {
      "epoch": 5.1572327044025155,
      "grad_norm": 0.30575695633888245,
      "learning_rate": 4.630069504337306e-05,
      "loss": 0.51,
      "num_input_tokens_seen": 6429760,
      "step": 9840
    },
    {
      "epoch": 5.159853249475891,
      "grad_norm": 0.38693055510520935,
      "learning_rate": 4.629470700707672e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 6432480,
      "step": 9845
    },
    {
      "epoch": 5.162473794549266,
      "grad_norm": 0.35005983710289,
      "learning_rate": 4.628871451617062e-05,
      "loss": 0.4411,
      "num_input_tokens_seen": 6437312,
      "step": 9850
    },
    {
      "epoch": 5.165094339622642,
      "grad_norm": 0.4312167167663574,
      "learning_rate": 4.628271757190832e-05,
      "loss": 0.5709,
      "num_input_tokens_seen": 6441664,
      "step": 9855
    },
    {
      "epoch": 5.167714884696017,
      "grad_norm": 0.2552902400493622,
      "learning_rate": 4.62767161755443e-05,
      "loss": 0.3881,
      "num_input_tokens_seen": 6445120,
      "step": 9860
    },
    {
      "epoch": 5.170335429769392,
      "grad_norm": 0.47043704986572266,
      "learning_rate": 4.6270710328334004e-05,
      "loss": 0.453,
      "num_input_tokens_seen": 6447968,
      "step": 9865
    },
    {
      "epoch": 5.172955974842767,
      "grad_norm": 0.5004812479019165,
      "learning_rate": 4.626470003153377e-05,
      "loss": 0.4542,
      "num_input_tokens_seen": 6450592,
      "step": 9870
    },
    {
      "epoch": 5.1755765199161425,
      "grad_norm": 0.692643404006958,
      "learning_rate": 4.625868528640089e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 6454496,
      "step": 9875
    },
    {
      "epoch": 5.178197064989518,
      "grad_norm": 0.4390397071838379,
      "learning_rate": 4.6252666094193586e-05,
      "loss": 0.3885,
      "num_input_tokens_seen": 6457088,
      "step": 9880
    },
    {
      "epoch": 5.180817610062893,
      "grad_norm": 0.37874388694763184,
      "learning_rate": 4.6246642456171e-05,
      "loss": 0.5511,
      "num_input_tokens_seen": 6461088,
      "step": 9885
    },
    {
      "epoch": 5.183438155136268,
      "grad_norm": 0.36477845907211304,
      "learning_rate": 4.6240614373593196e-05,
      "loss": 0.5084,
      "num_input_tokens_seen": 6463872,
      "step": 9890
    },
    {
      "epoch": 5.186058700209643,
      "grad_norm": 0.3942963182926178,
      "learning_rate": 4.623458184772119e-05,
      "loss": 0.4866,
      "num_input_tokens_seen": 6467424,
      "step": 9895
    },
    {
      "epoch": 5.188679245283019,
      "grad_norm": 0.5074275135993958,
      "learning_rate": 4.622854487981693e-05,
      "loss": 0.4435,
      "num_input_tokens_seen": 6470688,
      "step": 9900
    },
    {
      "epoch": 5.191299790356394,
      "grad_norm": 0.4112744927406311,
      "learning_rate": 4.622250347114326e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 6473440,
      "step": 9905
    },
    {
      "epoch": 5.19392033542977,
      "grad_norm": 0.7298882007598877,
      "learning_rate": 4.621645762296399e-05,
      "loss": 0.5522,
      "num_input_tokens_seen": 6476352,
      "step": 9910
    },
    {
      "epoch": 5.196540880503145,
      "grad_norm": 0.2973932921886444,
      "learning_rate": 4.6210407336543835e-05,
      "loss": 0.4028,
      "num_input_tokens_seen": 6480064,
      "step": 9915
    },
    {
      "epoch": 5.19916142557652,
      "grad_norm": 0.4237414002418518,
      "learning_rate": 4.620435261314844e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 6483840,
      "step": 9920
    },
    {
      "epoch": 5.201781970649895,
      "grad_norm": 0.49439704418182373,
      "learning_rate": 4.61982934540444e-05,
      "loss": 0.663,
      "num_input_tokens_seen": 6486368,
      "step": 9925
    },
    {
      "epoch": 5.20440251572327,
      "grad_norm": 0.28491267561912537,
      "learning_rate": 4.6192229860499206e-05,
      "loss": 0.4004,
      "num_input_tokens_seen": 6489152,
      "step": 9930
    },
    {
      "epoch": 5.2070230607966455,
      "grad_norm": 0.4327845871448517,
      "learning_rate": 4.6186161833781295e-05,
      "loss": 0.4421,
      "num_input_tokens_seen": 6493088,
      "step": 9935
    },
    {
      "epoch": 5.209643605870021,
      "grad_norm": 0.500106930732727,
      "learning_rate": 4.618008937516003e-05,
      "loss": 0.4401,
      "num_input_tokens_seen": 6496256,
      "step": 9940
    },
    {
      "epoch": 5.212264150943396,
      "grad_norm": 0.51055908203125,
      "learning_rate": 4.617401248590572e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 6499296,
      "step": 9945
    },
    {
      "epoch": 5.214884696016772,
      "grad_norm": 0.5392909049987793,
      "learning_rate": 4.616793116728955e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 6502624,
      "step": 9950
    },
    {
      "epoch": 5.217505241090147,
      "grad_norm": 0.32441169023513794,
      "learning_rate": 4.6161845420583685e-05,
      "loss": 0.658,
      "num_input_tokens_seen": 6506272,
      "step": 9955
    },
    {
      "epoch": 5.220125786163522,
      "grad_norm": 0.5362549424171448,
      "learning_rate": 4.6155755247061174e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 6509504,
      "step": 9960
    },
    {
      "epoch": 5.222746331236897,
      "grad_norm": 0.6669785976409912,
      "learning_rate": 4.614966064799603e-05,
      "loss": 0.5409,
      "num_input_tokens_seen": 6512512,
      "step": 9965
    },
    {
      "epoch": 5.2253668763102725,
      "grad_norm": 0.2300715297460556,
      "learning_rate": 4.6143561624663166e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 6515456,
      "step": 9970
    },
    {
      "epoch": 5.227987421383648,
      "grad_norm": 0.46287667751312256,
      "learning_rate": 4.6137458178338425e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 6518880,
      "step": 9975
    },
    {
      "epoch": 5.230607966457023,
      "grad_norm": 0.9101985692977905,
      "learning_rate": 4.613135031029859e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 6521664,
      "step": 9980
    },
    {
      "epoch": 5.233228511530398,
      "grad_norm": 0.5007807612419128,
      "learning_rate": 4.612523802182134e-05,
      "loss": 0.4638,
      "num_input_tokens_seen": 6524160,
      "step": 9985
    },
    {
      "epoch": 5.235849056603773,
      "grad_norm": 0.519006073474884,
      "learning_rate": 4.611912131418531e-05,
      "loss": 0.3735,
      "num_input_tokens_seen": 6527904,
      "step": 9990
    },
    {
      "epoch": 5.238469601677149,
      "grad_norm": 0.9444512128829956,
      "learning_rate": 4.611300018867004e-05,
      "loss": 0.7028,
      "num_input_tokens_seen": 6530560,
      "step": 9995
    },
    {
      "epoch": 5.241090146750524,
      "grad_norm": 0.5086673498153687,
      "learning_rate": 4.6106874646556e-05,
      "loss": 0.5046,
      "num_input_tokens_seen": 6534048,
      "step": 10000
    },
    {
      "epoch": 5.2437106918239,
      "grad_norm": 0.6196901798248291,
      "learning_rate": 4.6100744689124584e-05,
      "loss": 0.4328,
      "num_input_tokens_seen": 6536864,
      "step": 10005
    },
    {
      "epoch": 5.246331236897275,
      "grad_norm": 1.426705002784729,
      "learning_rate": 4.6094610317658116e-05,
      "loss": 0.5512,
      "num_input_tokens_seen": 6540608,
      "step": 10010
    },
    {
      "epoch": 5.24895178197065,
      "grad_norm": 0.5471377968788147,
      "learning_rate": 4.608847153343982e-05,
      "loss": 0.5209,
      "num_input_tokens_seen": 6543744,
      "step": 10015
    },
    {
      "epoch": 5.251572327044025,
      "grad_norm": 0.7186644077301025,
      "learning_rate": 4.6082328337753874e-05,
      "loss": 0.4348,
      "num_input_tokens_seen": 6546656,
      "step": 10020
    },
    {
      "epoch": 5.2541928721174,
      "grad_norm": 0.5686826705932617,
      "learning_rate": 4.6076180731885353e-05,
      "loss": 0.5255,
      "num_input_tokens_seen": 6549152,
      "step": 10025
    },
    {
      "epoch": 5.256813417190775,
      "grad_norm": 0.497368186712265,
      "learning_rate": 4.6070028717120286e-05,
      "loss": 0.5353,
      "num_input_tokens_seen": 6552096,
      "step": 10030
    },
    {
      "epoch": 5.259433962264151,
      "grad_norm": 0.2595456540584564,
      "learning_rate": 4.606387229474558e-05,
      "loss": 0.4916,
      "num_input_tokens_seen": 6555232,
      "step": 10035
    },
    {
      "epoch": 5.262054507337526,
      "grad_norm": 0.40360379219055176,
      "learning_rate": 4.60577114660491e-05,
      "loss": 0.417,
      "num_input_tokens_seen": 6557920,
      "step": 10040
    },
    {
      "epoch": 5.264675052410902,
      "grad_norm": 0.40138113498687744,
      "learning_rate": 4.605154623231962e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 6560928,
      "step": 10045
    },
    {
      "epoch": 5.267295597484277,
      "grad_norm": 0.41227805614471436,
      "learning_rate": 4.604537659484684e-05,
      "loss": 0.4291,
      "num_input_tokens_seen": 6563488,
      "step": 10050
    },
    {
      "epoch": 5.269916142557652,
      "grad_norm": 0.3542823791503906,
      "learning_rate": 4.603920255492138e-05,
      "loss": 0.4474,
      "num_input_tokens_seen": 6567520,
      "step": 10055
    },
    {
      "epoch": 5.272536687631027,
      "grad_norm": 0.3103925287723541,
      "learning_rate": 4.603302411383477e-05,
      "loss": 0.503,
      "num_input_tokens_seen": 6570528,
      "step": 10060
    },
    {
      "epoch": 5.2751572327044025,
      "grad_norm": 0.4761503040790558,
      "learning_rate": 4.602684127287946e-05,
      "loss": 0.4037,
      "num_input_tokens_seen": 6573824,
      "step": 10065
    },
    {
      "epoch": 5.277777777777778,
      "grad_norm": 0.37329015135765076,
      "learning_rate": 4.602065403334885e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 6576896,
      "step": 10070
    },
    {
      "epoch": 5.280398322851153,
      "grad_norm": 0.9447526335716248,
      "learning_rate": 4.601446239653724e-05,
      "loss": 0.4866,
      "num_input_tokens_seen": 6579136,
      "step": 10075
    },
    {
      "epoch": 5.283018867924528,
      "grad_norm": 0.39223766326904297,
      "learning_rate": 4.600826636373984e-05,
      "loss": 0.4181,
      "num_input_tokens_seen": 6582080,
      "step": 10080
    },
    {
      "epoch": 5.285639412997903,
      "grad_norm": 0.4626750349998474,
      "learning_rate": 4.600206593625279e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 6584896,
      "step": 10085
    },
    {
      "epoch": 5.288259958071279,
      "grad_norm": 0.49767160415649414,
      "learning_rate": 4.599586111537315e-05,
      "loss": 0.481,
      "num_input_tokens_seen": 6588448,
      "step": 10090
    },
    {
      "epoch": 5.290880503144654,
      "grad_norm": 0.5083041787147522,
      "learning_rate": 4.59896519023989e-05,
      "loss": 0.4007,
      "num_input_tokens_seen": 6592704,
      "step": 10095
    },
    {
      "epoch": 5.29350104821803,
      "grad_norm": 0.44683900475502014,
      "learning_rate": 4.5983438298628935e-05,
      "loss": 0.4559,
      "num_input_tokens_seen": 6596256,
      "step": 10100
    },
    {
      "epoch": 5.296121593291405,
      "grad_norm": 0.3905511200428009,
      "learning_rate": 4.597722030536306e-05,
      "loss": 0.5042,
      "num_input_tokens_seen": 6598656,
      "step": 10105
    },
    {
      "epoch": 5.29874213836478,
      "grad_norm": 0.34514859318733215,
      "learning_rate": 4.597099792390202e-05,
      "loss": 0.4871,
      "num_input_tokens_seen": 6601504,
      "step": 10110
    },
    {
      "epoch": 5.301362683438155,
      "grad_norm": 0.7315775752067566,
      "learning_rate": 4.596477115554748e-05,
      "loss": 0.5134,
      "num_input_tokens_seen": 6604640,
      "step": 10115
    },
    {
      "epoch": 5.30398322851153,
      "grad_norm": 0.3791198432445526,
      "learning_rate": 4.5958540001601976e-05,
      "loss": 0.5072,
      "num_input_tokens_seen": 6607520,
      "step": 10120
    },
    {
      "epoch": 5.306603773584905,
      "grad_norm": 0.35478129982948303,
      "learning_rate": 4.595230446336902e-05,
      "loss": 0.3898,
      "num_input_tokens_seen": 6611360,
      "step": 10125
    },
    {
      "epoch": 5.309224318658281,
      "grad_norm": 0.3190702795982361,
      "learning_rate": 4.5946064542153005e-05,
      "loss": 0.3684,
      "num_input_tokens_seen": 6616672,
      "step": 10130
    },
    {
      "epoch": 5.311844863731656,
      "grad_norm": 0.49325287342071533,
      "learning_rate": 4.593982023925926e-05,
      "loss": 0.6162,
      "num_input_tokens_seen": 6619424,
      "step": 10135
    },
    {
      "epoch": 5.314465408805032,
      "grad_norm": 0.6199206709861755,
      "learning_rate": 4.5933571555994e-05,
      "loss": 0.3333,
      "num_input_tokens_seen": 6622656,
      "step": 10140
    },
    {
      "epoch": 5.317085953878407,
      "grad_norm": 0.5049906373023987,
      "learning_rate": 4.59273184936644e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 6625568,
      "step": 10145
    },
    {
      "epoch": 5.319706498951782,
      "grad_norm": 1.1681571006774902,
      "learning_rate": 4.5921061053578523e-05,
      "loss": 0.6488,
      "num_input_tokens_seen": 6628896,
      "step": 10150
    },
    {
      "epoch": 5.322327044025157,
      "grad_norm": 0.587928295135498,
      "learning_rate": 4.5914799237045355e-05,
      "loss": 0.4236,
      "num_input_tokens_seen": 6632672,
      "step": 10155
    },
    {
      "epoch": 5.3249475890985325,
      "grad_norm": 0.4512779414653778,
      "learning_rate": 4.5908533045374796e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 6637344,
      "step": 10160
    },
    {
      "epoch": 5.327568134171908,
      "grad_norm": 0.7831221222877502,
      "learning_rate": 4.590226247987766e-05,
      "loss": 0.5798,
      "num_input_tokens_seen": 6640416,
      "step": 10165
    },
    {
      "epoch": 5.330188679245283,
      "grad_norm": 0.38174742460250854,
      "learning_rate": 4.589598754186568e-05,
      "loss": 0.3272,
      "num_input_tokens_seen": 6642688,
      "step": 10170
    },
    {
      "epoch": 5.332809224318658,
      "grad_norm": 0.6899113059043884,
      "learning_rate": 4.5889708232651494e-05,
      "loss": 0.4755,
      "num_input_tokens_seen": 6645952,
      "step": 10175
    },
    {
      "epoch": 5.335429769392033,
      "grad_norm": 0.5552715063095093,
      "learning_rate": 4.588342455354867e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 6648576,
      "step": 10180
    },
    {
      "epoch": 5.338050314465409,
      "grad_norm": 0.7067735195159912,
      "learning_rate": 4.5877136505871685e-05,
      "loss": 0.5528,
      "num_input_tokens_seen": 6651840,
      "step": 10185
    },
    {
      "epoch": 5.340670859538784,
      "grad_norm": 0.6697798371315002,
      "learning_rate": 4.5870844090935914e-05,
      "loss": 0.6216,
      "num_input_tokens_seen": 6654336,
      "step": 10190
    },
    {
      "epoch": 5.34329140461216,
      "grad_norm": 0.4163127839565277,
      "learning_rate": 4.586454731005768e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 6657600,
      "step": 10195
    },
    {
      "epoch": 5.345911949685535,
      "grad_norm": 0.3122755289077759,
      "learning_rate": 4.5858246164554175e-05,
      "loss": 0.4107,
      "num_input_tokens_seen": 6660768,
      "step": 10200
    },
    {
      "epoch": 5.34853249475891,
      "grad_norm": 0.8067290782928467,
      "learning_rate": 4.5851940655743536e-05,
      "loss": 0.5368,
      "num_input_tokens_seen": 6663808,
      "step": 10205
    },
    {
      "epoch": 5.351153039832285,
      "grad_norm": 0.2560417056083679,
      "learning_rate": 4.5845630784944804e-05,
      "loss": 0.4834,
      "num_input_tokens_seen": 6666048,
      "step": 10210
    },
    {
      "epoch": 5.35377358490566,
      "grad_norm": 0.3036472499370575,
      "learning_rate": 4.583931655347793e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 6668672,
      "step": 10215
    },
    {
      "epoch": 5.356394129979035,
      "grad_norm": 0.734318733215332,
      "learning_rate": 4.5832997962663785e-05,
      "loss": 0.6278,
      "num_input_tokens_seen": 6671680,
      "step": 10220
    },
    {
      "epoch": 5.359014675052411,
      "grad_norm": 0.4291570484638214,
      "learning_rate": 4.582667501382414e-05,
      "loss": 0.4024,
      "num_input_tokens_seen": 6674592,
      "step": 10225
    },
    {
      "epoch": 5.361635220125786,
      "grad_norm": 0.49754470586776733,
      "learning_rate": 4.582034770828169e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 6677632,
      "step": 10230
    },
    {
      "epoch": 5.364255765199162,
      "grad_norm": 0.7112358212471008,
      "learning_rate": 4.581401604736002e-05,
      "loss": 0.3943,
      "num_input_tokens_seen": 6680768,
      "step": 10235
    },
    {
      "epoch": 5.366876310272537,
      "grad_norm": 0.504570484161377,
      "learning_rate": 4.580768003238365e-05,
      "loss": 0.4999,
      "num_input_tokens_seen": 6683712,
      "step": 10240
    },
    {
      "epoch": 5.369496855345912,
      "grad_norm": 0.40196090936660767,
      "learning_rate": 4.5801339664678014e-05,
      "loss": 0.5238,
      "num_input_tokens_seen": 6688512,
      "step": 10245
    },
    {
      "epoch": 5.372117400419287,
      "grad_norm": 0.3441597521305084,
      "learning_rate": 4.579499494556943e-05,
      "loss": 0.438,
      "num_input_tokens_seen": 6691520,
      "step": 10250
    },
    {
      "epoch": 5.3747379454926625,
      "grad_norm": 0.2849646508693695,
      "learning_rate": 4.5788645876385147e-05,
      "loss": 0.5371,
      "num_input_tokens_seen": 6695360,
      "step": 10255
    },
    {
      "epoch": 5.377358490566038,
      "grad_norm": 0.2891318202018738,
      "learning_rate": 4.578229245845331e-05,
      "loss": 0.4607,
      "num_input_tokens_seen": 6698688,
      "step": 10260
    },
    {
      "epoch": 5.379979035639413,
      "grad_norm": 0.623226523399353,
      "learning_rate": 4.5775934693102994e-05,
      "loss": 0.5385,
      "num_input_tokens_seen": 6701312,
      "step": 10265
    },
    {
      "epoch": 5.382599580712788,
      "grad_norm": 0.4478963613510132,
      "learning_rate": 4.5769572581664165e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 6703840,
      "step": 10270
    },
    {
      "epoch": 5.385220125786163,
      "grad_norm": 0.3518502712249756,
      "learning_rate": 4.57632061254677e-05,
      "loss": 0.4231,
      "num_input_tokens_seen": 6706720,
      "step": 10275
    },
    {
      "epoch": 5.387840670859539,
      "grad_norm": 0.42075538635253906,
      "learning_rate": 4.5756835325845394e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 6710240,
      "step": 10280
    },
    {
      "epoch": 5.390461215932914,
      "grad_norm": 0.4789109230041504,
      "learning_rate": 4.575046018412995e-05,
      "loss": 0.5961,
      "num_input_tokens_seen": 6713632,
      "step": 10285
    },
    {
      "epoch": 5.3930817610062896,
      "grad_norm": 0.39494988322257996,
      "learning_rate": 4.5744080701654955e-05,
      "loss": 0.4204,
      "num_input_tokens_seen": 6716032,
      "step": 10290
    },
    {
      "epoch": 5.395702306079665,
      "grad_norm": 0.4224839210510254,
      "learning_rate": 4.573769687975495e-05,
      "loss": 0.4226,
      "num_input_tokens_seen": 6719136,
      "step": 10295
    },
    {
      "epoch": 5.39832285115304,
      "grad_norm": 0.663405179977417,
      "learning_rate": 4.573130871976534e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 6721824,
      "step": 10300
    },
    {
      "epoch": 5.400943396226415,
      "grad_norm": 0.32749325037002563,
      "learning_rate": 4.572491622302247e-05,
      "loss": 0.4594,
      "num_input_tokens_seen": 6724864,
      "step": 10305
    },
    {
      "epoch": 5.40356394129979,
      "grad_norm": 0.47221606969833374,
      "learning_rate": 4.571851939086356e-05,
      "loss": 0.3659,
      "num_input_tokens_seen": 6728064,
      "step": 10310
    },
    {
      "epoch": 5.406184486373165,
      "grad_norm": 0.4270828366279602,
      "learning_rate": 4.571211822462676e-05,
      "loss": 0.5321,
      "num_input_tokens_seen": 6730784,
      "step": 10315
    },
    {
      "epoch": 5.408805031446541,
      "grad_norm": 0.3554724156856537,
      "learning_rate": 4.570571272565114e-05,
      "loss": 0.5361,
      "num_input_tokens_seen": 6733408,
      "step": 10320
    },
    {
      "epoch": 5.411425576519916,
      "grad_norm": 0.05394595488905907,
      "learning_rate": 4.569930289527663e-05,
      "loss": 0.4633,
      "num_input_tokens_seen": 6738912,
      "step": 10325
    },
    {
      "epoch": 5.414046121593292,
      "grad_norm": 0.6283692717552185,
      "learning_rate": 4.56928887348441e-05,
      "loss": 0.4441,
      "num_input_tokens_seen": 6742784,
      "step": 10330
    },
    {
      "epoch": 5.416666666666667,
      "grad_norm": 0.3250088691711426,
      "learning_rate": 4.568647024569533e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 6746080,
      "step": 10335
    },
    {
      "epoch": 5.419287211740042,
      "grad_norm": 0.5427457690238953,
      "learning_rate": 4.568004742917299e-05,
      "loss": 0.5813,
      "num_input_tokens_seen": 6750368,
      "step": 10340
    },
    {
      "epoch": 5.421907756813417,
      "grad_norm": 0.5229172110557556,
      "learning_rate": 4.567362028662065e-05,
      "loss": 0.4525,
      "num_input_tokens_seen": 6753984,
      "step": 10345
    },
    {
      "epoch": 5.4245283018867925,
      "grad_norm": 0.39750614762306213,
      "learning_rate": 4.566718881938281e-05,
      "loss": 0.5705,
      "num_input_tokens_seen": 6756576,
      "step": 10350
    },
    {
      "epoch": 5.427148846960168,
      "grad_norm": 0.6373473405838013,
      "learning_rate": 4.5660753028804846e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 6759904,
      "step": 10355
    },
    {
      "epoch": 5.429769392033543,
      "grad_norm": 0.37700143456459045,
      "learning_rate": 4.5654312916233066e-05,
      "loss": 0.4492,
      "num_input_tokens_seen": 6762880,
      "step": 10360
    },
    {
      "epoch": 5.432389937106918,
      "grad_norm": 0.7084265351295471,
      "learning_rate": 4.564786848301466e-05,
      "loss": 0.6592,
      "num_input_tokens_seen": 6765728,
      "step": 10365
    },
    {
      "epoch": 5.435010482180293,
      "grad_norm": 0.9246441125869751,
      "learning_rate": 4.564141973049773e-05,
      "loss": 0.4328,
      "num_input_tokens_seen": 6768704,
      "step": 10370
    },
    {
      "epoch": 5.437631027253669,
      "grad_norm": 0.35924670100212097,
      "learning_rate": 4.563496666003128e-05,
      "loss": 0.3504,
      "num_input_tokens_seen": 6772256,
      "step": 10375
    },
    {
      "epoch": 5.440251572327044,
      "grad_norm": 0.3647141456604004,
      "learning_rate": 4.562850927296523e-05,
      "loss": 0.4925,
      "num_input_tokens_seen": 6775552,
      "step": 10380
    },
    {
      "epoch": 5.4428721174004195,
      "grad_norm": 0.5608070492744446,
      "learning_rate": 4.5622047570650365e-05,
      "loss": 0.3865,
      "num_input_tokens_seen": 6778176,
      "step": 10385
    },
    {
      "epoch": 5.445492662473795,
      "grad_norm": 0.3671855628490448,
      "learning_rate": 4.561558155443843e-05,
      "loss": 0.4622,
      "num_input_tokens_seen": 6781888,
      "step": 10390
    },
    {
      "epoch": 5.44811320754717,
      "grad_norm": 0.3668890595436096,
      "learning_rate": 4.560911122568202e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 6785056,
      "step": 10395
    },
    {
      "epoch": 5.450733752620545,
      "grad_norm": 0.6113609671592712,
      "learning_rate": 4.560263658573467e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 6788832,
      "step": 10400
    },
    {
      "epoch": 5.45335429769392,
      "grad_norm": 0.41361281275749207,
      "learning_rate": 4.559615763595079e-05,
      "loss": 0.4521,
      "num_input_tokens_seen": 6792320,
      "step": 10405
    },
    {
      "epoch": 5.455974842767295,
      "grad_norm": 0.44003888964653015,
      "learning_rate": 4.558967437768571e-05,
      "loss": 0.5683,
      "num_input_tokens_seen": 6796000,
      "step": 10410
    },
    {
      "epoch": 5.4585953878406706,
      "grad_norm": 0.6517558097839355,
      "learning_rate": 4.558318681229565e-05,
      "loss": 0.4563,
      "num_input_tokens_seen": 6798400,
      "step": 10415
    },
    {
      "epoch": 5.461215932914046,
      "grad_norm": 0.6148651242256165,
      "learning_rate": 4.557669494113773e-05,
      "loss": 0.5188,
      "num_input_tokens_seen": 6801440,
      "step": 10420
    },
    {
      "epoch": 5.463836477987422,
      "grad_norm": 0.5136642456054688,
      "learning_rate": 4.557019876556998e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 6807072,
      "step": 10425
    },
    {
      "epoch": 5.466457023060797,
      "grad_norm": 0.5977378487586975,
      "learning_rate": 4.556369828695133e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 6810272,
      "step": 10430
    },
    {
      "epoch": 5.469077568134172,
      "grad_norm": 0.48290300369262695,
      "learning_rate": 4.5557193506641596e-05,
      "loss": 0.4149,
      "num_input_tokens_seen": 6813184,
      "step": 10435
    },
    {
      "epoch": 5.471698113207547,
      "grad_norm": 0.4389157295227051,
      "learning_rate": 4.555068442600151e-05,
      "loss": 0.4682,
      "num_input_tokens_seen": 6817824,
      "step": 10440
    },
    {
      "epoch": 5.4743186582809225,
      "grad_norm": 0.4007461965084076,
      "learning_rate": 4.5544171046392703e-05,
      "loss": 0.4278,
      "num_input_tokens_seen": 6820480,
      "step": 10445
    },
    {
      "epoch": 5.476939203354298,
      "grad_norm": 0.8779600858688354,
      "learning_rate": 4.5537653369177685e-05,
      "loss": 0.5131,
      "num_input_tokens_seen": 6823360,
      "step": 10450
    },
    {
      "epoch": 5.479559748427673,
      "grad_norm": 0.45848995447158813,
      "learning_rate": 4.5531131395719895e-05,
      "loss": 0.499,
      "num_input_tokens_seen": 6825856,
      "step": 10455
    },
    {
      "epoch": 5.482180293501048,
      "grad_norm": 0.7542394399642944,
      "learning_rate": 4.552460512738364e-05,
      "loss": 0.5275,
      "num_input_tokens_seen": 6828896,
      "step": 10460
    },
    {
      "epoch": 5.484800838574423,
      "grad_norm": 0.441433310508728,
      "learning_rate": 4.551807456553416e-05,
      "loss": 0.5142,
      "num_input_tokens_seen": 6832032,
      "step": 10465
    },
    {
      "epoch": 5.487421383647799,
      "grad_norm": 0.7363870739936829,
      "learning_rate": 4.551153971153755e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 6834624,
      "step": 10470
    },
    {
      "epoch": 5.490041928721174,
      "grad_norm": 0.5150432586669922,
      "learning_rate": 4.550500056676085e-05,
      "loss": 0.6287,
      "num_input_tokens_seen": 6837504,
      "step": 10475
    },
    {
      "epoch": 5.4926624737945495,
      "grad_norm": 0.5755183696746826,
      "learning_rate": 4.5498457132571966e-05,
      "loss": 0.4935,
      "num_input_tokens_seen": 6840224,
      "step": 10480
    },
    {
      "epoch": 5.495283018867925,
      "grad_norm": 0.26218515634536743,
      "learning_rate": 4.5491909410339706e-05,
      "loss": 0.3612,
      "num_input_tokens_seen": 6843520,
      "step": 10485
    },
    {
      "epoch": 5.4979035639413,
      "grad_norm": 0.3002493977546692,
      "learning_rate": 4.548535740143378e-05,
      "loss": 0.4497,
      "num_input_tokens_seen": 6846464,
      "step": 10490
    },
    {
      "epoch": 5.500524109014675,
      "grad_norm": 0.5456089377403259,
      "learning_rate": 4.54788011072248e-05,
      "loss": 0.495,
      "num_input_tokens_seen": 6849696,
      "step": 10495
    },
    {
      "epoch": 5.50314465408805,
      "grad_norm": 0.4269234538078308,
      "learning_rate": 4.5472240529084254e-05,
      "loss": 0.6073,
      "num_input_tokens_seen": 6853216,
      "step": 10500
    },
    {
      "epoch": 5.505765199161425,
      "grad_norm": 0.5449066162109375,
      "learning_rate": 4.5465675668384557e-05,
      "loss": 0.4372,
      "num_input_tokens_seen": 6855872,
      "step": 10505
    },
    {
      "epoch": 5.5083857442348005,
      "grad_norm": 0.5776402950286865,
      "learning_rate": 4.545910652649899e-05,
      "loss": 0.394,
      "num_input_tokens_seen": 6858688,
      "step": 10510
    },
    {
      "epoch": 5.511006289308176,
      "grad_norm": 0.45925718545913696,
      "learning_rate": 4.5452533104801756e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 6861632,
      "step": 10515
    },
    {
      "epoch": 5.513626834381551,
      "grad_norm": 0.33040016889572144,
      "learning_rate": 4.544595540466792e-05,
      "loss": 0.6553,
      "num_input_tokens_seen": 6865696,
      "step": 10520
    },
    {
      "epoch": 5.516247379454927,
      "grad_norm": 0.5075066089630127,
      "learning_rate": 4.5439373427473476e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 6869056,
      "step": 10525
    },
    {
      "epoch": 5.518867924528302,
      "grad_norm": 0.9299365282058716,
      "learning_rate": 4.54327871745953e-05,
      "loss": 0.4403,
      "num_input_tokens_seen": 6871488,
      "step": 10530
    },
    {
      "epoch": 5.521488469601677,
      "grad_norm": 0.20356130599975586,
      "learning_rate": 4.542619664741116e-05,
      "loss": 0.4917,
      "num_input_tokens_seen": 6875008,
      "step": 10535
    },
    {
      "epoch": 5.524109014675052,
      "grad_norm": 0.7342448830604553,
      "learning_rate": 4.5419601847299706e-05,
      "loss": 0.6613,
      "num_input_tokens_seen": 6878336,
      "step": 10540
    },
    {
      "epoch": 5.526729559748428,
      "grad_norm": 0.3013931214809418,
      "learning_rate": 4.5413002775640503e-05,
      "loss": 0.4453,
      "num_input_tokens_seen": 6882080,
      "step": 10545
    },
    {
      "epoch": 5.529350104821803,
      "grad_norm": 0.323693186044693,
      "learning_rate": 4.540639943381401e-05,
      "loss": 0.3851,
      "num_input_tokens_seen": 6885504,
      "step": 10550
    },
    {
      "epoch": 5.531970649895178,
      "grad_norm": 0.3760283291339874,
      "learning_rate": 4.539979182320157e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 6888704,
      "step": 10555
    },
    {
      "epoch": 5.534591194968553,
      "grad_norm": 0.5512785911560059,
      "learning_rate": 4.5393179945185404e-05,
      "loss": 0.5463,
      "num_input_tokens_seen": 6891936,
      "step": 10560
    },
    {
      "epoch": 5.537211740041929,
      "grad_norm": 0.4504976272583008,
      "learning_rate": 4.538656380114865e-05,
      "loss": 0.5434,
      "num_input_tokens_seen": 6894848,
      "step": 10565
    },
    {
      "epoch": 5.539832285115304,
      "grad_norm": 0.5265846252441406,
      "learning_rate": 4.537994339247533e-05,
      "loss": 0.4602,
      "num_input_tokens_seen": 6897728,
      "step": 10570
    },
    {
      "epoch": 5.5424528301886795,
      "grad_norm": 0.6268934607505798,
      "learning_rate": 4.5373318720550366e-05,
      "loss": 0.4023,
      "num_input_tokens_seen": 6900672,
      "step": 10575
    },
    {
      "epoch": 5.545073375262055,
      "grad_norm": 0.44443944096565247,
      "learning_rate": 4.536668978675956e-05,
      "loss": 0.4235,
      "num_input_tokens_seen": 6905920,
      "step": 10580
    },
    {
      "epoch": 5.54769392033543,
      "grad_norm": 0.34079456329345703,
      "learning_rate": 4.53600565924896e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 6909408,
      "step": 10585
    },
    {
      "epoch": 5.550314465408805,
      "grad_norm": 0.5191816687583923,
      "learning_rate": 4.535341913912808e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 6912128,
      "step": 10590
    },
    {
      "epoch": 5.55293501048218,
      "grad_norm": 0.4796028733253479,
      "learning_rate": 4.534677742806348e-05,
      "loss": 0.5109,
      "num_input_tokens_seen": 6915136,
      "step": 10595
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 0.4447312355041504,
      "learning_rate": 4.5340131460685175e-05,
      "loss": 0.4654,
      "num_input_tokens_seen": 6917856,
      "step": 10600
    },
    {
      "epoch": 5.5581761006289305,
      "grad_norm": 0.445884108543396,
      "learning_rate": 4.533348123838341e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 6920864,
      "step": 10605
    },
    {
      "epoch": 5.560796645702306,
      "grad_norm": 0.4231657385826111,
      "learning_rate": 4.532682676254936e-05,
      "loss": 0.6595,
      "num_input_tokens_seen": 6924384,
      "step": 10610
    },
    {
      "epoch": 5.563417190775681,
      "grad_norm": 0.4877298176288605,
      "learning_rate": 4.5320168034575044e-05,
      "loss": 0.5014,
      "num_input_tokens_seen": 6926848,
      "step": 10615
    },
    {
      "epoch": 5.566037735849057,
      "grad_norm": 0.29365330934524536,
      "learning_rate": 4.53135050558534e-05,
      "loss": 0.3835,
      "num_input_tokens_seen": 6929536,
      "step": 10620
    },
    {
      "epoch": 5.568658280922432,
      "grad_norm": 0.6300483345985413,
      "learning_rate": 4.530683782777825e-05,
      "loss": 0.5328,
      "num_input_tokens_seen": 6932064,
      "step": 10625
    },
    {
      "epoch": 5.571278825995807,
      "grad_norm": 0.5107005834579468,
      "learning_rate": 4.53001663517443e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 6935296,
      "step": 10630
    },
    {
      "epoch": 5.573899371069182,
      "grad_norm": 0.5683975219726562,
      "learning_rate": 4.529349062914714e-05,
      "loss": 0.4126,
      "num_input_tokens_seen": 6937824,
      "step": 10635
    },
    {
      "epoch": 5.576519916142558,
      "grad_norm": 0.3949703276157379,
      "learning_rate": 4.528681066138326e-05,
      "loss": 0.4708,
      "num_input_tokens_seen": 6941600,
      "step": 10640
    },
    {
      "epoch": 5.579140461215933,
      "grad_norm": 0.3844149708747864,
      "learning_rate": 4.528012644985005e-05,
      "loss": 0.4637,
      "num_input_tokens_seen": 6944608,
      "step": 10645
    },
    {
      "epoch": 5.581761006289308,
      "grad_norm": 0.6165739297866821,
      "learning_rate": 4.527343799594574e-05,
      "loss": 0.5358,
      "num_input_tokens_seen": 6948000,
      "step": 10650
    },
    {
      "epoch": 5.584381551362683,
      "grad_norm": 0.339353084564209,
      "learning_rate": 4.5266745301069494e-05,
      "loss": 0.3653,
      "num_input_tokens_seen": 6951424,
      "step": 10655
    },
    {
      "epoch": 5.587002096436059,
      "grad_norm": 0.4005938470363617,
      "learning_rate": 4.526004836662135e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 6956000,
      "step": 10660
    },
    {
      "epoch": 5.589622641509434,
      "grad_norm": 0.47099247574806213,
      "learning_rate": 4.525334719400223e-05,
      "loss": 0.4757,
      "num_input_tokens_seen": 6958944,
      "step": 10665
    },
    {
      "epoch": 5.5922431865828095,
      "grad_norm": 0.24209833145141602,
      "learning_rate": 4.524664178461393e-05,
      "loss": 0.4148,
      "num_input_tokens_seen": 6962304,
      "step": 10670
    },
    {
      "epoch": 5.594863731656185,
      "grad_norm": 0.5198874473571777,
      "learning_rate": 4.5239932139859154e-05,
      "loss": 0.5692,
      "num_input_tokens_seen": 6965632,
      "step": 10675
    },
    {
      "epoch": 5.59748427672956,
      "grad_norm": 0.23290222883224487,
      "learning_rate": 4.523321826114149e-05,
      "loss": 0.5474,
      "num_input_tokens_seen": 6969312,
      "step": 10680
    },
    {
      "epoch": 5.600104821802935,
      "grad_norm": 0.31010374426841736,
      "learning_rate": 4.5226500149865395e-05,
      "loss": 0.3904,
      "num_input_tokens_seen": 6972576,
      "step": 10685
    },
    {
      "epoch": 5.60272536687631,
      "grad_norm": 0.41891345381736755,
      "learning_rate": 4.5219777807436224e-05,
      "loss": 0.3494,
      "num_input_tokens_seen": 6975872,
      "step": 10690
    },
    {
      "epoch": 5.605345911949685,
      "grad_norm": 0.2616357207298279,
      "learning_rate": 4.5213051235260225e-05,
      "loss": 0.3622,
      "num_input_tokens_seen": 6978784,
      "step": 10695
    },
    {
      "epoch": 5.6079664570230605,
      "grad_norm": 0.35169658064842224,
      "learning_rate": 4.52063204347445e-05,
      "loss": 0.6637,
      "num_input_tokens_seen": 6981408,
      "step": 10700
    },
    {
      "epoch": 5.610587002096436,
      "grad_norm": 0.42098379135131836,
      "learning_rate": 4.519958540729707e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 6984800,
      "step": 10705
    },
    {
      "epoch": 5.613207547169811,
      "grad_norm": 0.41954654455184937,
      "learning_rate": 4.519284615432682e-05,
      "loss": 0.4633,
      "num_input_tokens_seen": 6988192,
      "step": 10710
    },
    {
      "epoch": 5.615828092243187,
      "grad_norm": 0.3964773118495941,
      "learning_rate": 4.518610267724352e-05,
      "loss": 0.3792,
      "num_input_tokens_seen": 6992416,
      "step": 10715
    },
    {
      "epoch": 5.618448637316562,
      "grad_norm": 0.6630679965019226,
      "learning_rate": 4.517935497745786e-05,
      "loss": 0.411,
      "num_input_tokens_seen": 6995520,
      "step": 10720
    },
    {
      "epoch": 5.621069182389937,
      "grad_norm": 0.46328598260879517,
      "learning_rate": 4.517260305638133e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 6998464,
      "step": 10725
    },
    {
      "epoch": 5.623689727463312,
      "grad_norm": 0.4080456793308258,
      "learning_rate": 4.516584691542639e-05,
      "loss": 0.4407,
      "num_input_tokens_seen": 7001600,
      "step": 10730
    },
    {
      "epoch": 5.626310272536688,
      "grad_norm": 0.46405237913131714,
      "learning_rate": 4.515908655600634e-05,
      "loss": 0.481,
      "num_input_tokens_seen": 7004704,
      "step": 10735
    },
    {
      "epoch": 5.628930817610063,
      "grad_norm": 0.5178364515304565,
      "learning_rate": 4.515232197953537e-05,
      "loss": 0.5641,
      "num_input_tokens_seen": 7007424,
      "step": 10740
    },
    {
      "epoch": 5.631551362683438,
      "grad_norm": 0.42356351017951965,
      "learning_rate": 4.514555318742856e-05,
      "loss": 0.4085,
      "num_input_tokens_seen": 7010688,
      "step": 10745
    },
    {
      "epoch": 5.634171907756813,
      "grad_norm": 0.31536078453063965,
      "learning_rate": 4.513878018110184e-05,
      "loss": 0.3571,
      "num_input_tokens_seen": 7014048,
      "step": 10750
    },
    {
      "epoch": 5.636792452830189,
      "grad_norm": 0.5034300684928894,
      "learning_rate": 4.513200296197207e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 7017856,
      "step": 10755
    },
    {
      "epoch": 5.639412997903564,
      "grad_norm": 0.4070519208908081,
      "learning_rate": 4.512522153145696e-05,
      "loss": 0.3726,
      "num_input_tokens_seen": 7021248,
      "step": 10760
    },
    {
      "epoch": 5.6420335429769395,
      "grad_norm": 0.5605736374855042,
      "learning_rate": 4.5118435890975106e-05,
      "loss": 0.4568,
      "num_input_tokens_seen": 7024320,
      "step": 10765
    },
    {
      "epoch": 5.644654088050315,
      "grad_norm": 0.932479739189148,
      "learning_rate": 4.511164604194598e-05,
      "loss": 0.5127,
      "num_input_tokens_seen": 7027296,
      "step": 10770
    },
    {
      "epoch": 5.64727463312369,
      "grad_norm": 0.2754579782485962,
      "learning_rate": 4.5104851985789956e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 7030784,
      "step": 10775
    },
    {
      "epoch": 5.649895178197065,
      "grad_norm": 0.258799284696579,
      "learning_rate": 4.5098053723928254e-05,
      "loss": 0.676,
      "num_input_tokens_seen": 7034112,
      "step": 10780
    },
    {
      "epoch": 5.65251572327044,
      "grad_norm": 0.41800931096076965,
      "learning_rate": 4.509125125778302e-05,
      "loss": 0.3772,
      "num_input_tokens_seen": 7039072,
      "step": 10785
    },
    {
      "epoch": 5.655136268343815,
      "grad_norm": 0.5744259357452393,
      "learning_rate": 4.508444458877722e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 7042304,
      "step": 10790
    },
    {
      "epoch": 5.6577568134171905,
      "grad_norm": 0.36964601278305054,
      "learning_rate": 4.5077633718334755e-05,
      "loss": 0.5233,
      "num_input_tokens_seen": 7046400,
      "step": 10795
    },
    {
      "epoch": 5.660377358490566,
      "grad_norm": 0.43797630071640015,
      "learning_rate": 4.5070818647880375e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 7050656,
      "step": 10800
    },
    {
      "epoch": 5.662997903563941,
      "grad_norm": 0.4819944500923157,
      "learning_rate": 4.506399937883972e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 7053824,
      "step": 10805
    },
    {
      "epoch": 5.665618448637317,
      "grad_norm": 0.5417343974113464,
      "learning_rate": 4.5057175912639286e-05,
      "loss": 0.5261,
      "num_input_tokens_seen": 7057760,
      "step": 10810
    },
    {
      "epoch": 5.668238993710692,
      "grad_norm": 0.36257725954055786,
      "learning_rate": 4.5050348250706475e-05,
      "loss": 0.4035,
      "num_input_tokens_seen": 7061888,
      "step": 10815
    },
    {
      "epoch": 5.670859538784067,
      "grad_norm": 0.3635511100292206,
      "learning_rate": 4.5043516394469566e-05,
      "loss": 0.3551,
      "num_input_tokens_seen": 7064960,
      "step": 10820
    },
    {
      "epoch": 5.673480083857442,
      "grad_norm": 0.33195626735687256,
      "learning_rate": 4.503668034535768e-05,
      "loss": 0.4322,
      "num_input_tokens_seen": 7067968,
      "step": 10825
    },
    {
      "epoch": 5.676100628930818,
      "grad_norm": 0.26456576585769653,
      "learning_rate": 4.5029840104800876e-05,
      "loss": 0.5311,
      "num_input_tokens_seen": 7071424,
      "step": 10830
    },
    {
      "epoch": 5.678721174004193,
      "grad_norm": 0.5805847644805908,
      "learning_rate": 4.5022995674230025e-05,
      "loss": 0.5394,
      "num_input_tokens_seen": 7074656,
      "step": 10835
    },
    {
      "epoch": 5.681341719077568,
      "grad_norm": 0.5686337947845459,
      "learning_rate": 4.5016147055076914e-05,
      "loss": 0.4955,
      "num_input_tokens_seen": 7077952,
      "step": 10840
    },
    {
      "epoch": 5.683962264150943,
      "grad_norm": 0.44341376423835754,
      "learning_rate": 4.500929424877419e-05,
      "loss": 0.3517,
      "num_input_tokens_seen": 7080512,
      "step": 10845
    },
    {
      "epoch": 5.686582809224319,
      "grad_norm": 0.7091658115386963,
      "learning_rate": 4.5002437256755394e-05,
      "loss": 0.4226,
      "num_input_tokens_seen": 7083904,
      "step": 10850
    },
    {
      "epoch": 5.689203354297694,
      "grad_norm": 0.5131420493125916,
      "learning_rate": 4.4995576080454924e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 7086144,
      "step": 10855
    },
    {
      "epoch": 5.6918238993710695,
      "grad_norm": 0.34034329652786255,
      "learning_rate": 4.4988710721308056e-05,
      "loss": 0.4604,
      "num_input_tokens_seen": 7088608,
      "step": 10860
    },
    {
      "epoch": 5.694444444444445,
      "grad_norm": 0.2565823495388031,
      "learning_rate": 4.498184118075095e-05,
      "loss": 0.3979,
      "num_input_tokens_seen": 7091296,
      "step": 10865
    },
    {
      "epoch": 5.69706498951782,
      "grad_norm": 0.48729822039604187,
      "learning_rate": 4.497496746022063e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 7095008,
      "step": 10870
    },
    {
      "epoch": 5.699685534591195,
      "grad_norm": 0.6925845742225647,
      "learning_rate": 4.4968089561155015e-05,
      "loss": 0.4386,
      "num_input_tokens_seen": 7098016,
      "step": 10875
    },
    {
      "epoch": 5.70230607966457,
      "grad_norm": 0.31886765360832214,
      "learning_rate": 4.496120748499286e-05,
      "loss": 0.3925,
      "num_input_tokens_seen": 7101056,
      "step": 10880
    },
    {
      "epoch": 5.704926624737945,
      "grad_norm": 0.5736840963363647,
      "learning_rate": 4.4954321233173833e-05,
      "loss": 0.3811,
      "num_input_tokens_seen": 7103488,
      "step": 10885
    },
    {
      "epoch": 5.7075471698113205,
      "grad_norm": 0.3471297025680542,
      "learning_rate": 4.494743080713846e-05,
      "loss": 0.4035,
      "num_input_tokens_seen": 7107072,
      "step": 10890
    },
    {
      "epoch": 5.710167714884696,
      "grad_norm": 0.5953416228294373,
      "learning_rate": 4.494053620832812e-05,
      "loss": 0.4544,
      "num_input_tokens_seen": 7110336,
      "step": 10895
    },
    {
      "epoch": 5.712788259958071,
      "grad_norm": 0.4186568260192871,
      "learning_rate": 4.493363743818511e-05,
      "loss": 0.5433,
      "num_input_tokens_seen": 7114240,
      "step": 10900
    },
    {
      "epoch": 5.715408805031447,
      "grad_norm": 0.3184465765953064,
      "learning_rate": 4.492673449815257e-05,
      "loss": 0.5371,
      "num_input_tokens_seen": 7121216,
      "step": 10905
    },
    {
      "epoch": 5.718029350104822,
      "grad_norm": 0.8207269906997681,
      "learning_rate": 4.4919827389674496e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 7124800,
      "step": 10910
    },
    {
      "epoch": 5.720649895178197,
      "grad_norm": 0.5368495583534241,
      "learning_rate": 4.491291611419579e-05,
      "loss": 0.5202,
      "num_input_tokens_seen": 7128224,
      "step": 10915
    },
    {
      "epoch": 5.723270440251572,
      "grad_norm": 0.5523257255554199,
      "learning_rate": 4.4906000673162216e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 7131456,
      "step": 10920
    },
    {
      "epoch": 5.725890985324948,
      "grad_norm": 0.5149208307266235,
      "learning_rate": 4.48990810680204e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 7134944,
      "step": 10925
    },
    {
      "epoch": 5.728511530398323,
      "grad_norm": 0.24073447287082672,
      "learning_rate": 4.4892157300217844e-05,
      "loss": 0.5945,
      "num_input_tokens_seen": 7139200,
      "step": 10930
    },
    {
      "epoch": 5.731132075471698,
      "grad_norm": 0.5485483407974243,
      "learning_rate": 4.488522937120293e-05,
      "loss": 0.4237,
      "num_input_tokens_seen": 7141792,
      "step": 10935
    },
    {
      "epoch": 5.733752620545073,
      "grad_norm": 0.3353003263473511,
      "learning_rate": 4.4878297282424886e-05,
      "loss": 0.413,
      "num_input_tokens_seen": 7145152,
      "step": 10940
    },
    {
      "epoch": 5.736373165618449,
      "grad_norm": 0.4294361472129822,
      "learning_rate": 4.4871361035333836e-05,
      "loss": 0.585,
      "num_input_tokens_seen": 7148192,
      "step": 10945
    },
    {
      "epoch": 5.738993710691824,
      "grad_norm": 0.3894500136375427,
      "learning_rate": 4.486442063138076e-05,
      "loss": 0.3736,
      "num_input_tokens_seen": 7151520,
      "step": 10950
    },
    {
      "epoch": 5.7416142557651995,
      "grad_norm": 0.6047335863113403,
      "learning_rate": 4.4857476072017516e-05,
      "loss": 0.3752,
      "num_input_tokens_seen": 7154496,
      "step": 10955
    },
    {
      "epoch": 5.744234800838575,
      "grad_norm": 0.2968330681324005,
      "learning_rate": 4.485052735869682e-05,
      "loss": 0.5069,
      "num_input_tokens_seen": 7157600,
      "step": 10960
    },
    {
      "epoch": 5.74685534591195,
      "grad_norm": 0.5725052356719971,
      "learning_rate": 4.484357449287228e-05,
      "loss": 0.5366,
      "num_input_tokens_seen": 7160224,
      "step": 10965
    },
    {
      "epoch": 5.749475890985325,
      "grad_norm": 0.5070642828941345,
      "learning_rate": 4.483661747599833e-05,
      "loss": 0.3723,
      "num_input_tokens_seen": 7162720,
      "step": 10970
    },
    {
      "epoch": 5.7520964360587,
      "grad_norm": 0.7314313054084778,
      "learning_rate": 4.482965630953031e-05,
      "loss": 0.5735,
      "num_input_tokens_seen": 7166240,
      "step": 10975
    },
    {
      "epoch": 5.754716981132075,
      "grad_norm": 0.48944851756095886,
      "learning_rate": 4.4822690994924425e-05,
      "loss": 0.5284,
      "num_input_tokens_seen": 7172480,
      "step": 10980
    },
    {
      "epoch": 5.7573375262054505,
      "grad_norm": 0.3904643654823303,
      "learning_rate": 4.481572153363773e-05,
      "loss": 0.5296,
      "num_input_tokens_seen": 7176416,
      "step": 10985
    },
    {
      "epoch": 5.759958071278826,
      "grad_norm": 0.3496127128601074,
      "learning_rate": 4.480874792712817e-05,
      "loss": 0.5847,
      "num_input_tokens_seen": 7180832,
      "step": 10990
    },
    {
      "epoch": 5.762578616352201,
      "grad_norm": 0.5502843856811523,
      "learning_rate": 4.480177017685452e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 7184416,
      "step": 10995
    },
    {
      "epoch": 5.765199161425577,
      "grad_norm": 0.534412682056427,
      "learning_rate": 4.479478828427647e-05,
      "loss": 0.4247,
      "num_input_tokens_seen": 7186816,
      "step": 11000
    },
    {
      "epoch": 5.767819706498952,
      "grad_norm": 0.49755218625068665,
      "learning_rate": 4.478780225085454e-05,
      "loss": 0.5791,
      "num_input_tokens_seen": 7189824,
      "step": 11005
    },
    {
      "epoch": 5.770440251572327,
      "grad_norm": 0.5087746381759644,
      "learning_rate": 4.4780812078050125e-05,
      "loss": 0.512,
      "num_input_tokens_seen": 7192640,
      "step": 11010
    },
    {
      "epoch": 5.773060796645702,
      "grad_norm": 0.41222429275512695,
      "learning_rate": 4.47738177673255e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 7195328,
      "step": 11015
    },
    {
      "epoch": 5.7756813417190775,
      "grad_norm": 0.933539867401123,
      "learning_rate": 4.4766819320143785e-05,
      "loss": 0.6833,
      "num_input_tokens_seen": 7198464,
      "step": 11020
    },
    {
      "epoch": 5.778301886792453,
      "grad_norm": 0.2731834650039673,
      "learning_rate": 4.475981673796899e-05,
      "loss": 0.4538,
      "num_input_tokens_seen": 7202080,
      "step": 11025
    },
    {
      "epoch": 5.780922431865828,
      "grad_norm": 0.434650719165802,
      "learning_rate": 4.475281002226596e-05,
      "loss": 0.4828,
      "num_input_tokens_seen": 7204832,
      "step": 11030
    },
    {
      "epoch": 5.783542976939203,
      "grad_norm": 0.4963599443435669,
      "learning_rate": 4.4745799174500414e-05,
      "loss": 0.4003,
      "num_input_tokens_seen": 7208128,
      "step": 11035
    },
    {
      "epoch": 5.786163522012579,
      "grad_norm": 0.3937898576259613,
      "learning_rate": 4.473878419613896e-05,
      "loss": 0.4337,
      "num_input_tokens_seen": 7211232,
      "step": 11040
    },
    {
      "epoch": 5.788784067085954,
      "grad_norm": 0.35988733172416687,
      "learning_rate": 4.473176508864905e-05,
      "loss": 0.3974,
      "num_input_tokens_seen": 7214720,
      "step": 11045
    },
    {
      "epoch": 5.7914046121593294,
      "grad_norm": 0.5282117128372192,
      "learning_rate": 4.4724741853498995e-05,
      "loss": 0.5723,
      "num_input_tokens_seen": 7219200,
      "step": 11050
    },
    {
      "epoch": 5.794025157232705,
      "grad_norm": 0.321688175201416,
      "learning_rate": 4.471771449215797e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 7223808,
      "step": 11055
    },
    {
      "epoch": 5.79664570230608,
      "grad_norm": 0.8489549160003662,
      "learning_rate": 4.471068300609603e-05,
      "loss": 0.4994,
      "num_input_tokens_seen": 7227264,
      "step": 11060
    },
    {
      "epoch": 5.799266247379455,
      "grad_norm": 0.5448992848396301,
      "learning_rate": 4.4703647396784065e-05,
      "loss": 0.5548,
      "num_input_tokens_seen": 7232672,
      "step": 11065
    },
    {
      "epoch": 5.80188679245283,
      "grad_norm": 1.08509361743927,
      "learning_rate": 4.469660766569387e-05,
      "loss": 0.44,
      "num_input_tokens_seen": 7235584,
      "step": 11070
    },
    {
      "epoch": 5.804507337526205,
      "grad_norm": 0.44484397768974304,
      "learning_rate": 4.4689563814298057e-05,
      "loss": 0.5395,
      "num_input_tokens_seen": 7238592,
      "step": 11075
    },
    {
      "epoch": 5.8071278825995805,
      "grad_norm": 0.5008992552757263,
      "learning_rate": 4.468251584407012e-05,
      "loss": 0.4945,
      "num_input_tokens_seen": 7241952,
      "step": 11080
    },
    {
      "epoch": 5.809748427672956,
      "grad_norm": 0.4166201949119568,
      "learning_rate": 4.467546375648443e-05,
      "loss": 0.4301,
      "num_input_tokens_seen": 7245056,
      "step": 11085
    },
    {
      "epoch": 5.812368972746331,
      "grad_norm": 0.4976445436477661,
      "learning_rate": 4.4668407553016176e-05,
      "loss": 0.4637,
      "num_input_tokens_seen": 7248896,
      "step": 11090
    },
    {
      "epoch": 5.814989517819707,
      "grad_norm": 0.34269627928733826,
      "learning_rate": 4.466134723514146e-05,
      "loss": 0.5296,
      "num_input_tokens_seen": 7251680,
      "step": 11095
    },
    {
      "epoch": 5.817610062893082,
      "grad_norm": 0.40262946486473083,
      "learning_rate": 4.465428280433721e-05,
      "loss": 0.5806,
      "num_input_tokens_seen": 7254432,
      "step": 11100
    },
    {
      "epoch": 5.820230607966457,
      "grad_norm": 1.8894566297531128,
      "learning_rate": 4.464721426208123e-05,
      "loss": 0.5116,
      "num_input_tokens_seen": 7257280,
      "step": 11105
    },
    {
      "epoch": 5.822851153039832,
      "grad_norm": 0.33605918288230896,
      "learning_rate": 4.464014160985217e-05,
      "loss": 0.4268,
      "num_input_tokens_seen": 7259840,
      "step": 11110
    },
    {
      "epoch": 5.8254716981132075,
      "grad_norm": 0.5489059686660767,
      "learning_rate": 4.463306484912955e-05,
      "loss": 0.4458,
      "num_input_tokens_seen": 7262624,
      "step": 11115
    },
    {
      "epoch": 5.828092243186583,
      "grad_norm": 0.2966521680355072,
      "learning_rate": 4.462598398139375e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 7265568,
      "step": 11120
    },
    {
      "epoch": 5.830712788259958,
      "grad_norm": 0.6366423964500427,
      "learning_rate": 4.4618899008126005e-05,
      "loss": 0.3964,
      "num_input_tokens_seen": 7268896,
      "step": 11125
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 0.6037446856498718,
      "learning_rate": 4.461180993080841e-05,
      "loss": 0.3607,
      "num_input_tokens_seen": 7271328,
      "step": 11130
    },
    {
      "epoch": 5.835953878406709,
      "grad_norm": 0.5458166003227234,
      "learning_rate": 4.460471675092393e-05,
      "loss": 0.5568,
      "num_input_tokens_seen": 7274752,
      "step": 11135
    },
    {
      "epoch": 5.838574423480084,
      "grad_norm": 0.28915828466415405,
      "learning_rate": 4.4597619469956356e-05,
      "loss": 0.5622,
      "num_input_tokens_seen": 7279168,
      "step": 11140
    },
    {
      "epoch": 5.841194968553459,
      "grad_norm": 0.43250182271003723,
      "learning_rate": 4.459051808939038e-05,
      "loss": 0.4878,
      "num_input_tokens_seen": 7282944,
      "step": 11145
    },
    {
      "epoch": 5.843815513626835,
      "grad_norm": 0.5096742510795593,
      "learning_rate": 4.458341261071151e-05,
      "loss": 0.5139,
      "num_input_tokens_seen": 7285824,
      "step": 11150
    },
    {
      "epoch": 5.84643605870021,
      "grad_norm": 0.46510636806488037,
      "learning_rate": 4.457630303540615e-05,
      "loss": 0.4939,
      "num_input_tokens_seen": 7288384,
      "step": 11155
    },
    {
      "epoch": 5.849056603773585,
      "grad_norm": 0.4981638491153717,
      "learning_rate": 4.456918936496154e-05,
      "loss": 0.544,
      "num_input_tokens_seen": 7291840,
      "step": 11160
    },
    {
      "epoch": 5.85167714884696,
      "grad_norm": 0.45271509885787964,
      "learning_rate": 4.456207160086575e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 7296032,
      "step": 11165
    },
    {
      "epoch": 5.854297693920335,
      "grad_norm": 0.5821388363838196,
      "learning_rate": 4.4554949744607775e-05,
      "loss": 0.6119,
      "num_input_tokens_seen": 7298560,
      "step": 11170
    },
    {
      "epoch": 5.8569182389937104,
      "grad_norm": 0.40006908774375916,
      "learning_rate": 4.4547823797677405e-05,
      "loss": 0.3797,
      "num_input_tokens_seen": 7301376,
      "step": 11175
    },
    {
      "epoch": 5.859538784067086,
      "grad_norm": 0.611552357673645,
      "learning_rate": 4.45406937615653e-05,
      "loss": 0.4957,
      "num_input_tokens_seen": 7304896,
      "step": 11180
    },
    {
      "epoch": 5.862159329140461,
      "grad_norm": 0.3290058672428131,
      "learning_rate": 4.4533559637763006e-05,
      "loss": 0.4898,
      "num_input_tokens_seen": 7307968,
      "step": 11185
    },
    {
      "epoch": 5.864779874213837,
      "grad_norm": 0.5247215032577515,
      "learning_rate": 4.452642142776288e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 7311040,
      "step": 11190
    },
    {
      "epoch": 5.867400419287212,
      "grad_norm": 0.4601951241493225,
      "learning_rate": 4.451927913305816e-05,
      "loss": 0.4071,
      "num_input_tokens_seen": 7314848,
      "step": 11195
    },
    {
      "epoch": 5.870020964360587,
      "grad_norm": 0.4004581868648529,
      "learning_rate": 4.4512132755142935e-05,
      "loss": 0.5122,
      "num_input_tokens_seen": 7317440,
      "step": 11200
    },
    {
      "epoch": 5.872641509433962,
      "grad_norm": 0.4712711274623871,
      "learning_rate": 4.450498229551214e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 7320288,
      "step": 11205
    },
    {
      "epoch": 5.8752620545073375,
      "grad_norm": 0.5591554641723633,
      "learning_rate": 4.449782775566158e-05,
      "loss": 0.5204,
      "num_input_tokens_seen": 7323168,
      "step": 11210
    },
    {
      "epoch": 5.877882599580713,
      "grad_norm": 0.4456724524497986,
      "learning_rate": 4.449066913708789e-05,
      "loss": 0.4638,
      "num_input_tokens_seen": 7326048,
      "step": 11215
    },
    {
      "epoch": 5.880503144654088,
      "grad_norm": 0.46337369084358215,
      "learning_rate": 4.448350644128858e-05,
      "loss": 0.5127,
      "num_input_tokens_seen": 7328288,
      "step": 11220
    },
    {
      "epoch": 5.883123689727463,
      "grad_norm": 0.8723862171173096,
      "learning_rate": 4.4476339669762e-05,
      "loss": 0.5233,
      "num_input_tokens_seen": 7331328,
      "step": 11225
    },
    {
      "epoch": 5.885744234800838,
      "grad_norm": 0.4019998610019684,
      "learning_rate": 4.446916882400737e-05,
      "loss": 0.4394,
      "num_input_tokens_seen": 7334720,
      "step": 11230
    },
    {
      "epoch": 5.888364779874214,
      "grad_norm": 0.3097420930862427,
      "learning_rate": 4.446199390552472e-05,
      "loss": 0.3928,
      "num_input_tokens_seen": 7337312,
      "step": 11235
    },
    {
      "epoch": 5.890985324947589,
      "grad_norm": 0.4464956521987915,
      "learning_rate": 4.445481491581499e-05,
      "loss": 0.5158,
      "num_input_tokens_seen": 7340384,
      "step": 11240
    },
    {
      "epoch": 5.893605870020965,
      "grad_norm": 0.2700764536857605,
      "learning_rate": 4.444763185637993e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 7343936,
      "step": 11245
    },
    {
      "epoch": 5.89622641509434,
      "grad_norm": 0.48898372054100037,
      "learning_rate": 4.4440444728722166e-05,
      "loss": 0.4712,
      "num_input_tokens_seen": 7346848,
      "step": 11250
    },
    {
      "epoch": 5.898846960167715,
      "grad_norm": 0.42695045471191406,
      "learning_rate": 4.4433253534345136e-05,
      "loss": 0.512,
      "num_input_tokens_seen": 7350112,
      "step": 11255
    },
    {
      "epoch": 5.90146750524109,
      "grad_norm": 0.28604790568351746,
      "learning_rate": 4.442605827475318e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 7353952,
      "step": 11260
    },
    {
      "epoch": 5.904088050314465,
      "grad_norm": 0.7722171545028687,
      "learning_rate": 4.4418858951451457e-05,
      "loss": 0.5936,
      "num_input_tokens_seen": 7357440,
      "step": 11265
    },
    {
      "epoch": 5.90670859538784,
      "grad_norm": 0.40277519822120667,
      "learning_rate": 4.441165556594599e-05,
      "loss": 0.5934,
      "num_input_tokens_seen": 7360800,
      "step": 11270
    },
    {
      "epoch": 5.909329140461216,
      "grad_norm": 0.5751288533210754,
      "learning_rate": 4.440444811974363e-05,
      "loss": 0.5305,
      "num_input_tokens_seen": 7364032,
      "step": 11275
    },
    {
      "epoch": 5.911949685534591,
      "grad_norm": 0.5648027658462524,
      "learning_rate": 4.43972366143521e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 7366880,
      "step": 11280
    },
    {
      "epoch": 5.914570230607967,
      "grad_norm": 0.30047687888145447,
      "learning_rate": 4.4390021051279975e-05,
      "loss": 0.4197,
      "num_input_tokens_seen": 7369664,
      "step": 11285
    },
    {
      "epoch": 5.917190775681342,
      "grad_norm": 0.4275600016117096,
      "learning_rate": 4.438280143203665e-05,
      "loss": 0.5741,
      "num_input_tokens_seen": 7372192,
      "step": 11290
    },
    {
      "epoch": 5.919811320754717,
      "grad_norm": 0.7497261762619019,
      "learning_rate": 4.43755777581324e-05,
      "loss": 0.4991,
      "num_input_tokens_seen": 7374720,
      "step": 11295
    },
    {
      "epoch": 5.922431865828092,
      "grad_norm": 0.5791299939155579,
      "learning_rate": 4.4368350031078334e-05,
      "loss": 0.4743,
      "num_input_tokens_seen": 7378112,
      "step": 11300
    },
    {
      "epoch": 5.9250524109014675,
      "grad_norm": 0.461407870054245,
      "learning_rate": 4.43611182523864e-05,
      "loss": 0.454,
      "num_input_tokens_seen": 7381728,
      "step": 11305
    },
    {
      "epoch": 5.927672955974843,
      "grad_norm": 0.46640709042549133,
      "learning_rate": 4.4353882423569416e-05,
      "loss": 0.5009,
      "num_input_tokens_seen": 7385504,
      "step": 11310
    },
    {
      "epoch": 5.930293501048218,
      "grad_norm": 0.34650906920433044,
      "learning_rate": 4.434664254614103e-05,
      "loss": 0.3906,
      "num_input_tokens_seen": 7389600,
      "step": 11315
    },
    {
      "epoch": 5.932914046121593,
      "grad_norm": 0.5254861116409302,
      "learning_rate": 4.4339398621615734e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 7394240,
      "step": 11320
    },
    {
      "epoch": 5.935534591194968,
      "grad_norm": 0.32036322355270386,
      "learning_rate": 4.433215065150889e-05,
      "loss": 0.4577,
      "num_input_tokens_seen": 7396704,
      "step": 11325
    },
    {
      "epoch": 5.938155136268344,
      "grad_norm": 0.31626439094543457,
      "learning_rate": 4.4324898637336674e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 7400352,
      "step": 11330
    },
    {
      "epoch": 5.940775681341719,
      "grad_norm": 0.44195640087127686,
      "learning_rate": 4.431764258061615e-05,
      "loss": 0.3777,
      "num_input_tokens_seen": 7402816,
      "step": 11335
    },
    {
      "epoch": 5.943396226415095,
      "grad_norm": 0.5104120373725891,
      "learning_rate": 4.431038248286517e-05,
      "loss": 0.4547,
      "num_input_tokens_seen": 7405568,
      "step": 11340
    },
    {
      "epoch": 5.94601677148847,
      "grad_norm": 0.3822196424007416,
      "learning_rate": 4.430311834560248e-05,
      "loss": 0.4811,
      "num_input_tokens_seen": 7408608,
      "step": 11345
    },
    {
      "epoch": 5.948637316561845,
      "grad_norm": 0.6793649196624756,
      "learning_rate": 4.4295850170347664e-05,
      "loss": 0.5231,
      "num_input_tokens_seen": 7412448,
      "step": 11350
    },
    {
      "epoch": 5.95125786163522,
      "grad_norm": 0.6079685688018799,
      "learning_rate": 4.428857795862113e-05,
      "loss": 0.4123,
      "num_input_tokens_seen": 7415104,
      "step": 11355
    },
    {
      "epoch": 5.953878406708595,
      "grad_norm": 0.4987383484840393,
      "learning_rate": 4.428130171194414e-05,
      "loss": 0.4252,
      "num_input_tokens_seen": 7418944,
      "step": 11360
    },
    {
      "epoch": 5.95649895178197,
      "grad_norm": 0.2481360286474228,
      "learning_rate": 4.427402143183879e-05,
      "loss": 0.3914,
      "num_input_tokens_seen": 7422816,
      "step": 11365
    },
    {
      "epoch": 5.959119496855346,
      "grad_norm": 0.489054799079895,
      "learning_rate": 4.426673711982806e-05,
      "loss": 0.3924,
      "num_input_tokens_seen": 7425824,
      "step": 11370
    },
    {
      "epoch": 5.961740041928721,
      "grad_norm": 0.34075748920440674,
      "learning_rate": 4.425944877743573e-05,
      "loss": 0.5384,
      "num_input_tokens_seen": 7429088,
      "step": 11375
    },
    {
      "epoch": 5.964360587002097,
      "grad_norm": 0.3378289043903351,
      "learning_rate": 4.425215640618644e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 7432736,
      "step": 11380
    },
    {
      "epoch": 5.966981132075472,
      "grad_norm": 0.5368202924728394,
      "learning_rate": 4.424486000760567e-05,
      "loss": 0.5477,
      "num_input_tokens_seen": 7435840,
      "step": 11385
    },
    {
      "epoch": 5.969601677148847,
      "grad_norm": 0.6603004336357117,
      "learning_rate": 4.423755958321974e-05,
      "loss": 0.3968,
      "num_input_tokens_seen": 7441056,
      "step": 11390
    },
    {
      "epoch": 5.972222222222222,
      "grad_norm": 0.2656145691871643,
      "learning_rate": 4.423025513455582e-05,
      "loss": 0.4946,
      "num_input_tokens_seen": 7444640,
      "step": 11395
    },
    {
      "epoch": 5.9748427672955975,
      "grad_norm": 0.6266347169876099,
      "learning_rate": 4.422294666314192e-05,
      "loss": 0.5243,
      "num_input_tokens_seen": 7448096,
      "step": 11400
    },
    {
      "epoch": 5.977463312368973,
      "grad_norm": 0.40466850996017456,
      "learning_rate": 4.421563417050688e-05,
      "loss": 0.443,
      "num_input_tokens_seen": 7451520,
      "step": 11405
    },
    {
      "epoch": 5.980083857442348,
      "grad_norm": 0.5070726275444031,
      "learning_rate": 4.42083176581804e-05,
      "loss": 0.4207,
      "num_input_tokens_seen": 7454048,
      "step": 11410
    },
    {
      "epoch": 5.982704402515723,
      "grad_norm": 1.3499406576156616,
      "learning_rate": 4.4200997127693e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 7457216,
      "step": 11415
    },
    {
      "epoch": 5.985324947589098,
      "grad_norm": 0.3468188941478729,
      "learning_rate": 4.419367258057606e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 7460992,
      "step": 11420
    },
    {
      "epoch": 5.987945492662474,
      "grad_norm": 0.5668312311172485,
      "learning_rate": 4.418634401836179e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 7463488,
      "step": 11425
    },
    {
      "epoch": 5.990566037735849,
      "grad_norm": 0.3553057610988617,
      "learning_rate": 4.417901144258324e-05,
      "loss": 0.4703,
      "num_input_tokens_seen": 7466304,
      "step": 11430
    },
    {
      "epoch": 5.993186582809225,
      "grad_norm": 0.5375011563301086,
      "learning_rate": 4.4171674854774305e-05,
      "loss": 0.5178,
      "num_input_tokens_seen": 7470176,
      "step": 11435
    },
    {
      "epoch": 5.9958071278826,
      "grad_norm": 0.48825228214263916,
      "learning_rate": 4.416433425646971e-05,
      "loss": 0.3679,
      "num_input_tokens_seen": 7473376,
      "step": 11440
    },
    {
      "epoch": 5.998427672955975,
      "grad_norm": 0.7407269477844238,
      "learning_rate": 4.415698964920503e-05,
      "loss": 0.483,
      "num_input_tokens_seen": 7476864,
      "step": 11445
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.4840863347053528,
      "eval_runtime": 9.8238,
      "eval_samples_per_second": 86.321,
      "eval_steps_per_second": 21.58,
      "num_input_tokens_seen": 7478504,
      "step": 11448
    },
    {
      "epoch": 6.00104821802935,
      "grad_norm": 0.37178879976272583,
      "learning_rate": 4.414964103451667e-05,
      "loss": 0.4154,
      "num_input_tokens_seen": 7479688,
      "step": 11450
    },
    {
      "epoch": 6.003668763102725,
      "grad_norm": 0.37018370628356934,
      "learning_rate": 4.414228841394187e-05,
      "loss": 0.407,
      "num_input_tokens_seen": 7483176,
      "step": 11455
    },
    {
      "epoch": 6.0062893081761,
      "grad_norm": 0.3428807854652405,
      "learning_rate": 4.413493178901872e-05,
      "loss": 0.4849,
      "num_input_tokens_seen": 7486120,
      "step": 11460
    },
    {
      "epoch": 6.008909853249476,
      "grad_norm": 0.4543788731098175,
      "learning_rate": 4.412757116128614e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 7488872,
      "step": 11465
    },
    {
      "epoch": 6.011530398322851,
      "grad_norm": 0.7250048518180847,
      "learning_rate": 4.41202065322839e-05,
      "loss": 0.4981,
      "num_input_tokens_seen": 7491880,
      "step": 11470
    },
    {
      "epoch": 6.014150943396227,
      "grad_norm": 0.6804267168045044,
      "learning_rate": 4.411283790355259e-05,
      "loss": 0.3935,
      "num_input_tokens_seen": 7494952,
      "step": 11475
    },
    {
      "epoch": 6.016771488469602,
      "grad_norm": 0.2827664911746979,
      "learning_rate": 4.410546527663364e-05,
      "loss": 0.5164,
      "num_input_tokens_seen": 7498600,
      "step": 11480
    },
    {
      "epoch": 6.019392033542977,
      "grad_norm": 0.41779252886772156,
      "learning_rate": 4.409808865306931e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 7502696,
      "step": 11485
    },
    {
      "epoch": 6.022012578616352,
      "grad_norm": 0.40902194380760193,
      "learning_rate": 4.409070803440273e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 7505800,
      "step": 11490
    },
    {
      "epoch": 6.0246331236897275,
      "grad_norm": 0.665891170501709,
      "learning_rate": 4.4083323422177824e-05,
      "loss": 0.4657,
      "num_input_tokens_seen": 7508712,
      "step": 11495
    },
    {
      "epoch": 6.027253668763103,
      "grad_norm": 0.47491297125816345,
      "learning_rate": 4.407593481793937e-05,
      "loss": 0.38,
      "num_input_tokens_seen": 7511400,
      "step": 11500
    },
    {
      "epoch": 6.029874213836478,
      "grad_norm": 0.3501817584037781,
      "learning_rate": 4.406854222323298e-05,
      "loss": 0.4542,
      "num_input_tokens_seen": 7515208,
      "step": 11505
    },
    {
      "epoch": 6.032494758909853,
      "grad_norm": 0.5926520228385925,
      "learning_rate": 4.406114563960511e-05,
      "loss": 0.3012,
      "num_input_tokens_seen": 7520712,
      "step": 11510
    },
    {
      "epoch": 6.035115303983228,
      "grad_norm": 0.282610148191452,
      "learning_rate": 4.4053745068603026e-05,
      "loss": 0.4079,
      "num_input_tokens_seen": 7522984,
      "step": 11515
    },
    {
      "epoch": 6.037735849056604,
      "grad_norm": 0.2914521396160126,
      "learning_rate": 4.404634051177485e-05,
      "loss": 0.4569,
      "num_input_tokens_seen": 7526024,
      "step": 11520
    },
    {
      "epoch": 6.040356394129979,
      "grad_norm": 1.0182963609695435,
      "learning_rate": 4.403893197066953e-05,
      "loss": 0.5058,
      "num_input_tokens_seen": 7529288,
      "step": 11525
    },
    {
      "epoch": 6.0429769392033545,
      "grad_norm": 0.4230322539806366,
      "learning_rate": 4.4031519446836855e-05,
      "loss": 0.4441,
      "num_input_tokens_seen": 7531752,
      "step": 11530
    },
    {
      "epoch": 6.04559748427673,
      "grad_norm": 0.6024590134620667,
      "learning_rate": 4.402410294182743e-05,
      "loss": 0.4632,
      "num_input_tokens_seen": 7535112,
      "step": 11535
    },
    {
      "epoch": 6.048218029350105,
      "grad_norm": 0.5779078006744385,
      "learning_rate": 4.4016682457192714e-05,
      "loss": 0.4166,
      "num_input_tokens_seen": 7538344,
      "step": 11540
    },
    {
      "epoch": 6.05083857442348,
      "grad_norm": 0.8481634259223938,
      "learning_rate": 4.400925799448497e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 7541064,
      "step": 11545
    },
    {
      "epoch": 6.053459119496855,
      "grad_norm": 1.32985520362854,
      "learning_rate": 4.400182955525733e-05,
      "loss": 0.4199,
      "num_input_tokens_seen": 7544808,
      "step": 11550
    },
    {
      "epoch": 6.05607966457023,
      "grad_norm": 0.4347151219844818,
      "learning_rate": 4.3994397141063734e-05,
      "loss": 0.3518,
      "num_input_tokens_seen": 7548872,
      "step": 11555
    },
    {
      "epoch": 6.058700209643606,
      "grad_norm": 0.42836859822273254,
      "learning_rate": 4.398696075345896e-05,
      "loss": 0.5385,
      "num_input_tokens_seen": 7551912,
      "step": 11560
    },
    {
      "epoch": 6.061320754716981,
      "grad_norm": 0.2820272445678711,
      "learning_rate": 4.39795203939986e-05,
      "loss": 0.3824,
      "num_input_tokens_seen": 7554856,
      "step": 11565
    },
    {
      "epoch": 6.063941299790357,
      "grad_norm": 1.512273907661438,
      "learning_rate": 4.397207606423911e-05,
      "loss": 0.5596,
      "num_input_tokens_seen": 7558344,
      "step": 11570
    },
    {
      "epoch": 6.066561844863732,
      "grad_norm": 0.41681376099586487,
      "learning_rate": 4.396462776573775e-05,
      "loss": 0.4813,
      "num_input_tokens_seen": 7561704,
      "step": 11575
    },
    {
      "epoch": 6.069182389937107,
      "grad_norm": 0.4292794466018677,
      "learning_rate": 4.395717550005262e-05,
      "loss": 0.469,
      "num_input_tokens_seen": 7563816,
      "step": 11580
    },
    {
      "epoch": 6.071802935010482,
      "grad_norm": 0.3505373001098633,
      "learning_rate": 4.394971926874265e-05,
      "loss": 0.3976,
      "num_input_tokens_seen": 7566408,
      "step": 11585
    },
    {
      "epoch": 6.0744234800838575,
      "grad_norm": 0.5470808148384094,
      "learning_rate": 4.3942259073367604e-05,
      "loss": 0.44,
      "num_input_tokens_seen": 7569960,
      "step": 11590
    },
    {
      "epoch": 6.077044025157233,
      "grad_norm": 0.6820141673088074,
      "learning_rate": 4.3934794915488065e-05,
      "loss": 0.4085,
      "num_input_tokens_seen": 7572872,
      "step": 11595
    },
    {
      "epoch": 6.079664570230608,
      "grad_norm": 0.5254098773002625,
      "learning_rate": 4.392732679666544e-05,
      "loss": 0.587,
      "num_input_tokens_seen": 7576104,
      "step": 11600
    },
    {
      "epoch": 6.082285115303983,
      "grad_norm": 0.30632585287094116,
      "learning_rate": 4.3919854718461986e-05,
      "loss": 0.4075,
      "num_input_tokens_seen": 7578856,
      "step": 11605
    },
    {
      "epoch": 6.084905660377358,
      "grad_norm": 0.4221213757991791,
      "learning_rate": 4.391237868244078e-05,
      "loss": 0.4379,
      "num_input_tokens_seen": 7582408,
      "step": 11610
    },
    {
      "epoch": 6.087526205450734,
      "grad_norm": 0.3130001425743103,
      "learning_rate": 4.390489869016571e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 7585224,
      "step": 11615
    },
    {
      "epoch": 6.090146750524109,
      "grad_norm": 0.8375467658042908,
      "learning_rate": 4.389741474320152e-05,
      "loss": 0.4381,
      "num_input_tokens_seen": 7587880,
      "step": 11620
    },
    {
      "epoch": 6.0927672955974845,
      "grad_norm": 0.3469819724559784,
      "learning_rate": 4.388992684311376e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 7591784,
      "step": 11625
    },
    {
      "epoch": 6.09538784067086,
      "grad_norm": 0.7599107623100281,
      "learning_rate": 4.388243499146881e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 7594920,
      "step": 11630
    },
    {
      "epoch": 6.098008385744235,
      "grad_norm": 0.6687374114990234,
      "learning_rate": 4.3874939189833876e-05,
      "loss": 0.5872,
      "num_input_tokens_seen": 7598504,
      "step": 11635
    },
    {
      "epoch": 6.10062893081761,
      "grad_norm": 0.31578415632247925,
      "learning_rate": 4.3867439439777005e-05,
      "loss": 0.5332,
      "num_input_tokens_seen": 7601640,
      "step": 11640
    },
    {
      "epoch": 6.103249475890985,
      "grad_norm": 0.562110960483551,
      "learning_rate": 4.385993574286705e-05,
      "loss": 0.4941,
      "num_input_tokens_seen": 7604520,
      "step": 11645
    },
    {
      "epoch": 6.10587002096436,
      "grad_norm": 0.23122195899486542,
      "learning_rate": 4.3852428100673705e-05,
      "loss": 0.3342,
      "num_input_tokens_seen": 7608136,
      "step": 11650
    },
    {
      "epoch": 6.1084905660377355,
      "grad_norm": 0.5211713314056396,
      "learning_rate": 4.3844916514767484e-05,
      "loss": 0.5175,
      "num_input_tokens_seen": 7611592,
      "step": 11655
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 0.9676951169967651,
      "learning_rate": 4.383740098671972e-05,
      "loss": 0.4685,
      "num_input_tokens_seen": 7614184,
      "step": 11660
    },
    {
      "epoch": 6.113731656184487,
      "grad_norm": 0.5399523377418518,
      "learning_rate": 4.3829881518102576e-05,
      "loss": 0.3886,
      "num_input_tokens_seen": 7617064,
      "step": 11665
    },
    {
      "epoch": 6.116352201257862,
      "grad_norm": 0.3782190978527069,
      "learning_rate": 4.3822358110489045e-05,
      "loss": 0.4009,
      "num_input_tokens_seen": 7620680,
      "step": 11670
    },
    {
      "epoch": 6.118972746331237,
      "grad_norm": 0.4037776589393616,
      "learning_rate": 4.381483076545294e-05,
      "loss": 0.4133,
      "num_input_tokens_seen": 7623944,
      "step": 11675
    },
    {
      "epoch": 6.121593291404612,
      "grad_norm": 0.7704584002494812,
      "learning_rate": 4.380729948456888e-05,
      "loss": 0.4179,
      "num_input_tokens_seen": 7627144,
      "step": 11680
    },
    {
      "epoch": 6.1242138364779874,
      "grad_norm": 0.39393869042396545,
      "learning_rate": 4.3799764269412336e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 7630216,
      "step": 11685
    },
    {
      "epoch": 6.126834381551363,
      "grad_norm": 0.6457173824310303,
      "learning_rate": 4.379222512155958e-05,
      "loss": 0.4641,
      "num_input_tokens_seen": 7633000,
      "step": 11690
    },
    {
      "epoch": 6.129454926624738,
      "grad_norm": 0.7243542671203613,
      "learning_rate": 4.378468204258773e-05,
      "loss": 0.3833,
      "num_input_tokens_seen": 7636104,
      "step": 11695
    },
    {
      "epoch": 6.132075471698113,
      "grad_norm": 0.5172021985054016,
      "learning_rate": 4.37771350340747e-05,
      "loss": 0.5976,
      "num_input_tokens_seen": 7639208,
      "step": 11700
    },
    {
      "epoch": 6.134696016771488,
      "grad_norm": 0.3148917555809021,
      "learning_rate": 4.376958409759925e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 7642184,
      "step": 11705
    },
    {
      "epoch": 6.137316561844864,
      "grad_norm": 0.8871679902076721,
      "learning_rate": 4.376202923474093e-05,
      "loss": 0.4864,
      "num_input_tokens_seen": 7645480,
      "step": 11710
    },
    {
      "epoch": 6.139937106918239,
      "grad_norm": 0.27603524923324585,
      "learning_rate": 4.3754470447080145e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 7648936,
      "step": 11715
    },
    {
      "epoch": 6.1425576519916145,
      "grad_norm": 0.507353663444519,
      "learning_rate": 4.374690773619811e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 7651816,
      "step": 11720
    },
    {
      "epoch": 6.14517819706499,
      "grad_norm": 0.29639390110969543,
      "learning_rate": 4.373934110367685e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 7654568,
      "step": 11725
    },
    {
      "epoch": 6.147798742138365,
      "grad_norm": 0.37744542956352234,
      "learning_rate": 4.373177055109922e-05,
      "loss": 0.2959,
      "num_input_tokens_seen": 7657544,
      "step": 11730
    },
    {
      "epoch": 6.15041928721174,
      "grad_norm": 0.5554347038269043,
      "learning_rate": 4.3724196080048896e-05,
      "loss": 0.5252,
      "num_input_tokens_seen": 7660936,
      "step": 11735
    },
    {
      "epoch": 6.153039832285115,
      "grad_norm": 0.557268500328064,
      "learning_rate": 4.3716617692110364e-05,
      "loss": 0.5053,
      "num_input_tokens_seen": 7663336,
      "step": 11740
    },
    {
      "epoch": 6.15566037735849,
      "grad_norm": 0.37102413177490234,
      "learning_rate": 4.3709035388868943e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 7666312,
      "step": 11745
    },
    {
      "epoch": 6.1582809224318655,
      "grad_norm": 0.5427832007408142,
      "learning_rate": 4.3701449171910767e-05,
      "loss": 0.5272,
      "num_input_tokens_seen": 7669224,
      "step": 11750
    },
    {
      "epoch": 6.160901467505241,
      "grad_norm": 0.3182979226112366,
      "learning_rate": 4.3693859042822774e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 7672264,
      "step": 11755
    },
    {
      "epoch": 6.163522012578617,
      "grad_norm": 0.49299269914627075,
      "learning_rate": 4.368626500319275e-05,
      "loss": 0.5292,
      "num_input_tokens_seen": 7674824,
      "step": 11760
    },
    {
      "epoch": 6.166142557651992,
      "grad_norm": 0.6462395787239075,
      "learning_rate": 4.367866705460927e-05,
      "loss": 0.333,
      "num_input_tokens_seen": 7677608,
      "step": 11765
    },
    {
      "epoch": 6.168763102725367,
      "grad_norm": 0.5614611506462097,
      "learning_rate": 4.367106519866173e-05,
      "loss": 0.4854,
      "num_input_tokens_seen": 7680680,
      "step": 11770
    },
    {
      "epoch": 6.171383647798742,
      "grad_norm": 0.4142756462097168,
      "learning_rate": 4.366345943694037e-05,
      "loss": 0.4809,
      "num_input_tokens_seen": 7684072,
      "step": 11775
    },
    {
      "epoch": 6.174004192872117,
      "grad_norm": 0.2615634500980377,
      "learning_rate": 4.365584977103622e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 7687624,
      "step": 11780
    },
    {
      "epoch": 6.176624737945493,
      "grad_norm": 0.3047195374965668,
      "learning_rate": 4.364823620254113e-05,
      "loss": 0.5485,
      "num_input_tokens_seen": 7691656,
      "step": 11785
    },
    {
      "epoch": 6.179245283018868,
      "grad_norm": 0.3128828704357147,
      "learning_rate": 4.364061873304778e-05,
      "loss": 0.5018,
      "num_input_tokens_seen": 7694568,
      "step": 11790
    },
    {
      "epoch": 6.181865828092243,
      "grad_norm": 0.19493699073791504,
      "learning_rate": 4.363299736414965e-05,
      "loss": 0.3861,
      "num_input_tokens_seen": 7697544,
      "step": 11795
    },
    {
      "epoch": 6.184486373165618,
      "grad_norm": 0.4109989106655121,
      "learning_rate": 4.3625372097441045e-05,
      "loss": 0.7478,
      "num_input_tokens_seen": 7700456,
      "step": 11800
    },
    {
      "epoch": 6.187106918238993,
      "grad_norm": 0.532804012298584,
      "learning_rate": 4.3617742934517094e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 7703112,
      "step": 11805
    },
    {
      "epoch": 6.189727463312369,
      "grad_norm": 0.3985055088996887,
      "learning_rate": 4.361010987697373e-05,
      "loss": 0.4177,
      "num_input_tokens_seen": 7705992,
      "step": 11810
    },
    {
      "epoch": 6.1923480083857445,
      "grad_norm": 0.42924344539642334,
      "learning_rate": 4.360247292640768e-05,
      "loss": 0.4685,
      "num_input_tokens_seen": 7709576,
      "step": 11815
    },
    {
      "epoch": 6.19496855345912,
      "grad_norm": 0.41594022512435913,
      "learning_rate": 4.359483208441653e-05,
      "loss": 0.4466,
      "num_input_tokens_seen": 7712712,
      "step": 11820
    },
    {
      "epoch": 6.197589098532495,
      "grad_norm": 0.7335275411605835,
      "learning_rate": 4.3587187352598646e-05,
      "loss": 0.4371,
      "num_input_tokens_seen": 7715432,
      "step": 11825
    },
    {
      "epoch": 6.20020964360587,
      "grad_norm": 0.3640820384025574,
      "learning_rate": 4.357953873255323e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 7718216,
      "step": 11830
    },
    {
      "epoch": 6.202830188679245,
      "grad_norm": 0.44373971223831177,
      "learning_rate": 4.357188622588027e-05,
      "loss": 0.4286,
      "num_input_tokens_seen": 7721096,
      "step": 11835
    },
    {
      "epoch": 6.20545073375262,
      "grad_norm": 0.6477983593940735,
      "learning_rate": 4.356422983418059e-05,
      "loss": 0.5425,
      "num_input_tokens_seen": 7724264,
      "step": 11840
    },
    {
      "epoch": 6.2080712788259955,
      "grad_norm": 0.5054333209991455,
      "learning_rate": 4.3556569559055824e-05,
      "loss": 0.4174,
      "num_input_tokens_seen": 7727176,
      "step": 11845
    },
    {
      "epoch": 6.210691823899371,
      "grad_norm": 0.5721979737281799,
      "learning_rate": 4.3548905402108406e-05,
      "loss": 0.4729,
      "num_input_tokens_seen": 7729544,
      "step": 11850
    },
    {
      "epoch": 6.213312368972747,
      "grad_norm": 0.3458622395992279,
      "learning_rate": 4.35412373649416e-05,
      "loss": 0.5409,
      "num_input_tokens_seen": 7733288,
      "step": 11855
    },
    {
      "epoch": 6.215932914046122,
      "grad_norm": 0.44124823808670044,
      "learning_rate": 4.353356544915947e-05,
      "loss": 0.4025,
      "num_input_tokens_seen": 7737096,
      "step": 11860
    },
    {
      "epoch": 6.218553459119497,
      "grad_norm": 0.7900480628013611,
      "learning_rate": 4.352588965636689e-05,
      "loss": 0.4445,
      "num_input_tokens_seen": 7741512,
      "step": 11865
    },
    {
      "epoch": 6.221174004192872,
      "grad_norm": 0.47933676838874817,
      "learning_rate": 4.351820998816954e-05,
      "loss": 0.4534,
      "num_input_tokens_seen": 7744648,
      "step": 11870
    },
    {
      "epoch": 6.223794549266247,
      "grad_norm": 0.9365295171737671,
      "learning_rate": 4.3510526446173935e-05,
      "loss": 0.3732,
      "num_input_tokens_seen": 7748392,
      "step": 11875
    },
    {
      "epoch": 6.226415094339623,
      "grad_norm": 0.5759475827217102,
      "learning_rate": 4.350283903198737e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 7752360,
      "step": 11880
    },
    {
      "epoch": 6.229035639412998,
      "grad_norm": 0.5909157395362854,
      "learning_rate": 4.349514774721797e-05,
      "loss": 0.3261,
      "num_input_tokens_seen": 7755016,
      "step": 11885
    },
    {
      "epoch": 6.231656184486373,
      "grad_norm": 0.3661949932575226,
      "learning_rate": 4.348745259347467e-05,
      "loss": 0.5095,
      "num_input_tokens_seen": 7758120,
      "step": 11890
    },
    {
      "epoch": 6.234276729559748,
      "grad_norm": 0.7103599309921265,
      "learning_rate": 4.347975357236721e-05,
      "loss": 0.6185,
      "num_input_tokens_seen": 7761032,
      "step": 11895
    },
    {
      "epoch": 6.236897274633123,
      "grad_norm": 0.4734673500061035,
      "learning_rate": 4.347205068550612e-05,
      "loss": 0.4177,
      "num_input_tokens_seen": 7763240,
      "step": 11900
    },
    {
      "epoch": 6.239517819706499,
      "grad_norm": 0.31530892848968506,
      "learning_rate": 4.3464343934502774e-05,
      "loss": 0.4473,
      "num_input_tokens_seen": 7766664,
      "step": 11905
    },
    {
      "epoch": 6.2421383647798745,
      "grad_norm": 0.5650310516357422,
      "learning_rate": 4.345663332096932e-05,
      "loss": 0.5506,
      "num_input_tokens_seen": 7769448,
      "step": 11910
    },
    {
      "epoch": 6.24475890985325,
      "grad_norm": 0.9623122811317444,
      "learning_rate": 4.344891884651875e-05,
      "loss": 0.5649,
      "num_input_tokens_seen": 7772296,
      "step": 11915
    },
    {
      "epoch": 6.247379454926625,
      "grad_norm": 0.5953649878501892,
      "learning_rate": 4.344120051276483e-05,
      "loss": 0.3831,
      "num_input_tokens_seen": 7775912,
      "step": 11920
    },
    {
      "epoch": 6.25,
      "grad_norm": 0.3879886567592621,
      "learning_rate": 4.3433478321322155e-05,
      "loss": 0.522,
      "num_input_tokens_seen": 7779784,
      "step": 11925
    },
    {
      "epoch": 6.252620545073375,
      "grad_norm": 0.3664211630821228,
      "learning_rate": 4.3425752273806106e-05,
      "loss": 0.3429,
      "num_input_tokens_seen": 7782504,
      "step": 11930
    },
    {
      "epoch": 6.25524109014675,
      "grad_norm": 0.657141923904419,
      "learning_rate": 4.34180223718329e-05,
      "loss": 0.3277,
      "num_input_tokens_seen": 7785864,
      "step": 11935
    },
    {
      "epoch": 6.2578616352201255,
      "grad_norm": 0.6152829527854919,
      "learning_rate": 4.341028861701953e-05,
      "loss": 0.3992,
      "num_input_tokens_seen": 7788808,
      "step": 11940
    },
    {
      "epoch": 6.260482180293501,
      "grad_norm": 0.576971709728241,
      "learning_rate": 4.3402551010983825e-05,
      "loss": 0.4222,
      "num_input_tokens_seen": 7792232,
      "step": 11945
    },
    {
      "epoch": 6.263102725366876,
      "grad_norm": 0.352090060710907,
      "learning_rate": 4.33948095553444e-05,
      "loss": 0.4276,
      "num_input_tokens_seen": 7794856,
      "step": 11950
    },
    {
      "epoch": 6.265723270440252,
      "grad_norm": 0.41918662190437317,
      "learning_rate": 4.338706425172067e-05,
      "loss": 0.5853,
      "num_input_tokens_seen": 7797800,
      "step": 11955
    },
    {
      "epoch": 6.268343815513627,
      "grad_norm": 0.28747549653053284,
      "learning_rate": 4.3379315101732866e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 7801256,
      "step": 11960
    },
    {
      "epoch": 6.270964360587002,
      "grad_norm": 0.33622848987579346,
      "learning_rate": 4.3371562107002026e-05,
      "loss": 0.3954,
      "num_input_tokens_seen": 7804296,
      "step": 11965
    },
    {
      "epoch": 6.273584905660377,
      "grad_norm": 0.49497538805007935,
      "learning_rate": 4.3363805269149984e-05,
      "loss": 0.4738,
      "num_input_tokens_seen": 7807464,
      "step": 11970
    },
    {
      "epoch": 6.276205450733753,
      "grad_norm": 0.21574613451957703,
      "learning_rate": 4.335604458979939e-05,
      "loss": 0.4219,
      "num_input_tokens_seen": 7812008,
      "step": 11975
    },
    {
      "epoch": 6.278825995807128,
      "grad_norm": 0.8876062631607056,
      "learning_rate": 4.3348280070573686e-05,
      "loss": 0.4548,
      "num_input_tokens_seen": 7815336,
      "step": 11980
    },
    {
      "epoch": 6.281446540880503,
      "grad_norm": 0.25171971321105957,
      "learning_rate": 4.334051171309712e-05,
      "loss": 0.3929,
      "num_input_tokens_seen": 7818856,
      "step": 11985
    },
    {
      "epoch": 6.284067085953878,
      "grad_norm": 0.4246448576450348,
      "learning_rate": 4.333273951899474e-05,
      "loss": 0.5572,
      "num_input_tokens_seen": 7822024,
      "step": 11990
    },
    {
      "epoch": 6.286687631027253,
      "grad_norm": 0.3971071243286133,
      "learning_rate": 4.33249634898924e-05,
      "loss": 0.613,
      "num_input_tokens_seen": 7825640,
      "step": 11995
    },
    {
      "epoch": 6.289308176100629,
      "grad_norm": 0.3286978602409363,
      "learning_rate": 4.331718362741677e-05,
      "loss": 0.4372,
      "num_input_tokens_seen": 7827944,
      "step": 12000
    },
    {
      "epoch": 6.2919287211740045,
      "grad_norm": 0.4371315836906433,
      "learning_rate": 4.330939993319529e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 7830600,
      "step": 12005
    },
    {
      "epoch": 6.29454926624738,
      "grad_norm": 0.27474334836006165,
      "learning_rate": 4.330161240885623e-05,
      "loss": 0.3501,
      "num_input_tokens_seen": 7834056,
      "step": 12010
    },
    {
      "epoch": 6.297169811320755,
      "grad_norm": 0.4801557660102844,
      "learning_rate": 4.3293821056028654e-05,
      "loss": 0.444,
      "num_input_tokens_seen": 7836904,
      "step": 12015
    },
    {
      "epoch": 6.29979035639413,
      "grad_norm": 0.7220028042793274,
      "learning_rate": 4.3286025876342415e-05,
      "loss": 0.467,
      "num_input_tokens_seen": 7839880,
      "step": 12020
    },
    {
      "epoch": 6.302410901467505,
      "grad_norm": 0.3563905954360962,
      "learning_rate": 4.327822687142819e-05,
      "loss": 0.4307,
      "num_input_tokens_seen": 7843432,
      "step": 12025
    },
    {
      "epoch": 6.30503144654088,
      "grad_norm": 0.49164390563964844,
      "learning_rate": 4.327042404291742e-05,
      "loss": 0.4155,
      "num_input_tokens_seen": 7846216,
      "step": 12030
    },
    {
      "epoch": 6.3076519916142555,
      "grad_norm": 0.6542354226112366,
      "learning_rate": 4.326261739244239e-05,
      "loss": 0.4701,
      "num_input_tokens_seen": 7849544,
      "step": 12035
    },
    {
      "epoch": 6.310272536687631,
      "grad_norm": 0.5313556790351868,
      "learning_rate": 4.325480692163615e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 7852552,
      "step": 12040
    },
    {
      "epoch": 6.312893081761006,
      "grad_norm": 0.5278922319412231,
      "learning_rate": 4.324699263213256e-05,
      "loss": 0.5435,
      "num_input_tokens_seen": 7855336,
      "step": 12045
    },
    {
      "epoch": 6.315513626834382,
      "grad_norm": 0.34209516644477844,
      "learning_rate": 4.3239174525566285e-05,
      "loss": 0.5046,
      "num_input_tokens_seen": 7858184,
      "step": 12050
    },
    {
      "epoch": 6.318134171907757,
      "grad_norm": 0.68893963098526,
      "learning_rate": 4.3231352603572784e-05,
      "loss": 0.3853,
      "num_input_tokens_seen": 7861192,
      "step": 12055
    },
    {
      "epoch": 6.320754716981132,
      "grad_norm": 0.3823937773704529,
      "learning_rate": 4.322352686778832e-05,
      "loss": 0.4494,
      "num_input_tokens_seen": 7864520,
      "step": 12060
    },
    {
      "epoch": 6.323375262054507,
      "grad_norm": 0.6596480011940002,
      "learning_rate": 4.3215697319849926e-05,
      "loss": 0.5377,
      "num_input_tokens_seen": 7867240,
      "step": 12065
    },
    {
      "epoch": 6.325995807127883,
      "grad_norm": 0.6331934332847595,
      "learning_rate": 4.3207863961395476e-05,
      "loss": 0.5145,
      "num_input_tokens_seen": 7870120,
      "step": 12070
    },
    {
      "epoch": 6.328616352201258,
      "grad_norm": 0.7555115222930908,
      "learning_rate": 4.320002679406361e-05,
      "loss": 0.3633,
      "num_input_tokens_seen": 7873736,
      "step": 12075
    },
    {
      "epoch": 6.331236897274633,
      "grad_norm": 0.5234948992729187,
      "learning_rate": 4.319218581949378e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 7877000,
      "step": 12080
    },
    {
      "epoch": 6.333857442348008,
      "grad_norm": 0.5612115859985352,
      "learning_rate": 4.318434103932622e-05,
      "loss": 0.3389,
      "num_input_tokens_seen": 7879784,
      "step": 12085
    },
    {
      "epoch": 6.336477987421383,
      "grad_norm": 0.5029840469360352,
      "learning_rate": 4.317649245520197e-05,
      "loss": 0.5266,
      "num_input_tokens_seen": 7882792,
      "step": 12090
    },
    {
      "epoch": 6.339098532494759,
      "grad_norm": 0.6758719682693481,
      "learning_rate": 4.3168640068762875e-05,
      "loss": 0.5147,
      "num_input_tokens_seen": 7885416,
      "step": 12095
    },
    {
      "epoch": 6.3417190775681345,
      "grad_norm": 0.3355991840362549,
      "learning_rate": 4.316078388165155e-05,
      "loss": 0.4882,
      "num_input_tokens_seen": 7888936,
      "step": 12100
    },
    {
      "epoch": 6.34433962264151,
      "grad_norm": 0.27450212836265564,
      "learning_rate": 4.315292389551143e-05,
      "loss": 0.5227,
      "num_input_tokens_seen": 7891976,
      "step": 12105
    },
    {
      "epoch": 6.346960167714885,
      "grad_norm": 0.7961565852165222,
      "learning_rate": 4.3145060111986726e-05,
      "loss": 0.5542,
      "num_input_tokens_seen": 7895848,
      "step": 12110
    },
    {
      "epoch": 6.34958071278826,
      "grad_norm": 0.4844578802585602,
      "learning_rate": 4.313719253272246e-05,
      "loss": 0.6783,
      "num_input_tokens_seen": 7898760,
      "step": 12115
    },
    {
      "epoch": 6.352201257861635,
      "grad_norm": 0.6457654237747192,
      "learning_rate": 4.3129321159364436e-05,
      "loss": 0.4461,
      "num_input_tokens_seen": 7902120,
      "step": 12120
    },
    {
      "epoch": 6.35482180293501,
      "grad_norm": 0.984446108341217,
      "learning_rate": 4.312144599355925e-05,
      "loss": 0.6521,
      "num_input_tokens_seen": 7905224,
      "step": 12125
    },
    {
      "epoch": 6.3574423480083855,
      "grad_norm": 0.46221092343330383,
      "learning_rate": 4.31135670369543e-05,
      "loss": 0.5633,
      "num_input_tokens_seen": 7908424,
      "step": 12130
    },
    {
      "epoch": 6.360062893081761,
      "grad_norm": 0.5082088112831116,
      "learning_rate": 4.310568429119779e-05,
      "loss": 0.4489,
      "num_input_tokens_seen": 7911336,
      "step": 12135
    },
    {
      "epoch": 6.362683438155136,
      "grad_norm": 0.40074849128723145,
      "learning_rate": 4.309779775793868e-05,
      "loss": 0.3603,
      "num_input_tokens_seen": 7913704,
      "step": 12140
    },
    {
      "epoch": 6.365303983228512,
      "grad_norm": 0.725980818271637,
      "learning_rate": 4.308990743882675e-05,
      "loss": 0.5198,
      "num_input_tokens_seen": 7916584,
      "step": 12145
    },
    {
      "epoch": 6.367924528301887,
      "grad_norm": 0.3407323360443115,
      "learning_rate": 4.308201333551256e-05,
      "loss": 0.5592,
      "num_input_tokens_seen": 7919880,
      "step": 12150
    },
    {
      "epoch": 6.370545073375262,
      "grad_norm": 0.5422291159629822,
      "learning_rate": 4.3074115449647476e-05,
      "loss": 0.5289,
      "num_input_tokens_seen": 7923752,
      "step": 12155
    },
    {
      "epoch": 6.373165618448637,
      "grad_norm": 0.42427846789360046,
      "learning_rate": 4.306621378288364e-05,
      "loss": 0.5536,
      "num_input_tokens_seen": 7927464,
      "step": 12160
    },
    {
      "epoch": 6.3757861635220126,
      "grad_norm": 0.5192249417304993,
      "learning_rate": 4.305830833687399e-05,
      "loss": 0.375,
      "num_input_tokens_seen": 7930792,
      "step": 12165
    },
    {
      "epoch": 6.378406708595388,
      "grad_norm": 0.4169484078884125,
      "learning_rate": 4.305039911327225e-05,
      "loss": 0.4881,
      "num_input_tokens_seen": 7935112,
      "step": 12170
    },
    {
      "epoch": 6.381027253668763,
      "grad_norm": 0.37933558225631714,
      "learning_rate": 4.304248611373296e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 7938152,
      "step": 12175
    },
    {
      "epoch": 6.383647798742138,
      "grad_norm": 0.3730257451534271,
      "learning_rate": 4.30345693399114e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 7942536,
      "step": 12180
    },
    {
      "epoch": 6.386268343815513,
      "grad_norm": 0.49390408396720886,
      "learning_rate": 4.30266487934637e-05,
      "loss": 0.4009,
      "num_input_tokens_seen": 7945224,
      "step": 12185
    },
    {
      "epoch": 6.388888888888889,
      "grad_norm": 0.5314028859138489,
      "learning_rate": 4.3018724476046714e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 7948520,
      "step": 12190
    },
    {
      "epoch": 6.3915094339622645,
      "grad_norm": 0.6310353875160217,
      "learning_rate": 4.301079638931814e-05,
      "loss": 0.5238,
      "num_input_tokens_seen": 7951368,
      "step": 12195
    },
    {
      "epoch": 6.39412997903564,
      "grad_norm": 0.3586427569389343,
      "learning_rate": 4.3002864534936444e-05,
      "loss": 0.6089,
      "num_input_tokens_seen": 7954472,
      "step": 12200
    },
    {
      "epoch": 6.396750524109015,
      "grad_norm": 0.42149174213409424,
      "learning_rate": 4.2994928914560874e-05,
      "loss": 0.5112,
      "num_input_tokens_seen": 7957000,
      "step": 12205
    },
    {
      "epoch": 6.39937106918239,
      "grad_norm": 0.44979748129844666,
      "learning_rate": 4.298698952985147e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 7959400,
      "step": 12210
    },
    {
      "epoch": 6.401991614255765,
      "grad_norm": 1.1345412731170654,
      "learning_rate": 4.297904638246906e-05,
      "loss": 0.4692,
      "num_input_tokens_seen": 7962312,
      "step": 12215
    },
    {
      "epoch": 6.40461215932914,
      "grad_norm": 0.6327104568481445,
      "learning_rate": 4.297109947407527e-05,
      "loss": 0.5458,
      "num_input_tokens_seen": 7965768,
      "step": 12220
    },
    {
      "epoch": 6.4072327044025155,
      "grad_norm": 0.3930809795856476,
      "learning_rate": 4.2963148806332486e-05,
      "loss": 0.3974,
      "num_input_tokens_seen": 7972840,
      "step": 12225
    },
    {
      "epoch": 6.409853249475891,
      "grad_norm": 0.47694066166877747,
      "learning_rate": 4.2955194380903915e-05,
      "loss": 0.5288,
      "num_input_tokens_seen": 7975304,
      "step": 12230
    },
    {
      "epoch": 6.412473794549266,
      "grad_norm": 0.48487940430641174,
      "learning_rate": 4.2947236199453524e-05,
      "loss": 0.3858,
      "num_input_tokens_seen": 7977928,
      "step": 12235
    },
    {
      "epoch": 6.415094339622642,
      "grad_norm": 0.47678983211517334,
      "learning_rate": 4.2939274263646065e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 7982120,
      "step": 12240
    },
    {
      "epoch": 6.417714884696017,
      "grad_norm": 0.4827331006526947,
      "learning_rate": 4.29313085751471e-05,
      "loss": 0.4436,
      "num_input_tokens_seen": 7985032,
      "step": 12245
    },
    {
      "epoch": 6.420335429769392,
      "grad_norm": 0.5198033452033997,
      "learning_rate": 4.292333913562294e-05,
      "loss": 0.4374,
      "num_input_tokens_seen": 7987272,
      "step": 12250
    },
    {
      "epoch": 6.422955974842767,
      "grad_norm": 0.4379611611366272,
      "learning_rate": 4.291536594674073e-05,
      "loss": 0.5058,
      "num_input_tokens_seen": 7990152,
      "step": 12255
    },
    {
      "epoch": 6.4255765199161425,
      "grad_norm": 0.42214855551719666,
      "learning_rate": 4.2907389010168344e-05,
      "loss": 0.5307,
      "num_input_tokens_seen": 7992648,
      "step": 12260
    },
    {
      "epoch": 6.428197064989518,
      "grad_norm": 0.33288267254829407,
      "learning_rate": 4.2899408327574475e-05,
      "loss": 0.4361,
      "num_input_tokens_seen": 7995880,
      "step": 12265
    },
    {
      "epoch": 6.430817610062893,
      "grad_norm": 1.109492540359497,
      "learning_rate": 4.289142390062859e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 7998312,
      "step": 12270
    },
    {
      "epoch": 6.433438155136268,
      "grad_norm": 0.3659663796424866,
      "learning_rate": 4.2883435731000946e-05,
      "loss": 0.4349,
      "num_input_tokens_seen": 8001768,
      "step": 12275
    },
    {
      "epoch": 6.436058700209643,
      "grad_norm": 1.0789443254470825,
      "learning_rate": 4.287544382036256e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 8004392,
      "step": 12280
    },
    {
      "epoch": 6.438679245283019,
      "grad_norm": 0.38052430748939514,
      "learning_rate": 4.286744817038527e-05,
      "loss": 0.4362,
      "num_input_tokens_seen": 8006984,
      "step": 12285
    },
    {
      "epoch": 6.441299790356394,
      "grad_norm": 0.3306204080581665,
      "learning_rate": 4.2859448782741666e-05,
      "loss": 0.6181,
      "num_input_tokens_seen": 8010920,
      "step": 12290
    },
    {
      "epoch": 6.44392033542977,
      "grad_norm": 0.47402772307395935,
      "learning_rate": 4.2851445659105126e-05,
      "loss": 0.5226,
      "num_input_tokens_seen": 8013800,
      "step": 12295
    },
    {
      "epoch": 6.446540880503145,
      "grad_norm": 0.6013581156730652,
      "learning_rate": 4.284343880114982e-05,
      "loss": 0.3771,
      "num_input_tokens_seen": 8016616,
      "step": 12300
    },
    {
      "epoch": 6.44916142557652,
      "grad_norm": 0.25703248381614685,
      "learning_rate": 4.2835428210550676e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 8020040,
      "step": 12305
    },
    {
      "epoch": 6.451781970649895,
      "grad_norm": 0.38347503542900085,
      "learning_rate": 4.282741388898344e-05,
      "loss": 0.3802,
      "num_input_tokens_seen": 8024008,
      "step": 12310
    },
    {
      "epoch": 6.45440251572327,
      "grad_norm": 0.44554343819618225,
      "learning_rate": 4.281939583812459e-05,
      "loss": 0.4881,
      "num_input_tokens_seen": 8027528,
      "step": 12315
    },
    {
      "epoch": 6.4570230607966455,
      "grad_norm": 1.1588462591171265,
      "learning_rate": 4.281137405965143e-05,
      "loss": 0.5385,
      "num_input_tokens_seen": 8030312,
      "step": 12320
    },
    {
      "epoch": 6.459643605870021,
      "grad_norm": 0.375272274017334,
      "learning_rate": 4.280334855524203e-05,
      "loss": 0.3477,
      "num_input_tokens_seen": 8034280,
      "step": 12325
    },
    {
      "epoch": 6.462264150943396,
      "grad_norm": 0.360495388507843,
      "learning_rate": 4.279531932657521e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 8038408,
      "step": 12330
    },
    {
      "epoch": 6.464884696016772,
      "grad_norm": 0.5137851238250732,
      "learning_rate": 4.278728637533062e-05,
      "loss": 0.4592,
      "num_input_tokens_seen": 8041544,
      "step": 12335
    },
    {
      "epoch": 6.467505241090147,
      "grad_norm": 0.2786681652069092,
      "learning_rate": 4.2779249703188626e-05,
      "loss": 0.3146,
      "num_input_tokens_seen": 8045096,
      "step": 12340
    },
    {
      "epoch": 6.470125786163522,
      "grad_norm": 0.43152573704719543,
      "learning_rate": 4.277120931183045e-05,
      "loss": 0.4162,
      "num_input_tokens_seen": 8049064,
      "step": 12345
    },
    {
      "epoch": 6.472746331236897,
      "grad_norm": 0.4380820095539093,
      "learning_rate": 4.276316520293803e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 8051816,
      "step": 12350
    },
    {
      "epoch": 6.4753668763102725,
      "grad_norm": 0.5080118179321289,
      "learning_rate": 4.27551173781941e-05,
      "loss": 0.4657,
      "num_input_tokens_seen": 8055528,
      "step": 12355
    },
    {
      "epoch": 6.477987421383648,
      "grad_norm": 0.33368751406669617,
      "learning_rate": 4.274706583928216e-05,
      "loss": 0.3903,
      "num_input_tokens_seen": 8058472,
      "step": 12360
    },
    {
      "epoch": 6.480607966457023,
      "grad_norm": 0.4823896586894989,
      "learning_rate": 4.273901058788654e-05,
      "loss": 0.6847,
      "num_input_tokens_seen": 8061608,
      "step": 12365
    },
    {
      "epoch": 6.483228511530398,
      "grad_norm": 0.2443103939294815,
      "learning_rate": 4.2730951625692275e-05,
      "loss": 0.4803,
      "num_input_tokens_seen": 8064648,
      "step": 12370
    },
    {
      "epoch": 6.485849056603773,
      "grad_norm": 0.4684642255306244,
      "learning_rate": 4.27228889543852e-05,
      "loss": 0.6339,
      "num_input_tokens_seen": 8067496,
      "step": 12375
    },
    {
      "epoch": 6.488469601677149,
      "grad_norm": 0.3349064886569977,
      "learning_rate": 4.2714822575651966e-05,
      "loss": 0.4026,
      "num_input_tokens_seen": 8069864,
      "step": 12380
    },
    {
      "epoch": 6.491090146750524,
      "grad_norm": 0.49514567852020264,
      "learning_rate": 4.270675249117994e-05,
      "loss": 0.5715,
      "num_input_tokens_seen": 8072936,
      "step": 12385
    },
    {
      "epoch": 6.4937106918239,
      "grad_norm": 0.7754908204078674,
      "learning_rate": 4.269867870265731e-05,
      "loss": 0.5383,
      "num_input_tokens_seen": 8075784,
      "step": 12390
    },
    {
      "epoch": 6.496331236897275,
      "grad_norm": 0.926723301410675,
      "learning_rate": 4.269060121177301e-05,
      "loss": 0.4986,
      "num_input_tokens_seen": 8078280,
      "step": 12395
    },
    {
      "epoch": 6.49895178197065,
      "grad_norm": 0.36523911356925964,
      "learning_rate": 4.2682520020216764e-05,
      "loss": 0.3582,
      "num_input_tokens_seen": 8081992,
      "step": 12400
    },
    {
      "epoch": 6.501572327044025,
      "grad_norm": 0.49128469824790955,
      "learning_rate": 4.267443512967906e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 8085672,
      "step": 12405
    },
    {
      "epoch": 6.5041928721174,
      "grad_norm": 0.2523767352104187,
      "learning_rate": 4.2666346541851165e-05,
      "loss": 0.396,
      "num_input_tokens_seen": 8089000,
      "step": 12410
    },
    {
      "epoch": 6.506813417190775,
      "grad_norm": 0.3391929268836975,
      "learning_rate": 4.265825425842512e-05,
      "loss": 0.4026,
      "num_input_tokens_seen": 8092168,
      "step": 12415
    },
    {
      "epoch": 6.509433962264151,
      "grad_norm": 0.3717588484287262,
      "learning_rate": 4.265015828109375e-05,
      "loss": 0.4961,
      "num_input_tokens_seen": 8094728,
      "step": 12420
    },
    {
      "epoch": 6.512054507337526,
      "grad_norm": 1.1609268188476562,
      "learning_rate": 4.264205861155063e-05,
      "loss": 0.5056,
      "num_input_tokens_seen": 8098088,
      "step": 12425
    },
    {
      "epoch": 6.514675052410902,
      "grad_norm": 0.5842971205711365,
      "learning_rate": 4.263395525149011e-05,
      "loss": 0.4701,
      "num_input_tokens_seen": 8100904,
      "step": 12430
    },
    {
      "epoch": 6.517295597484277,
      "grad_norm": 0.47552579641342163,
      "learning_rate": 4.262584820260734e-05,
      "loss": 0.383,
      "num_input_tokens_seen": 8104264,
      "step": 12435
    },
    {
      "epoch": 6.519916142557652,
      "grad_norm": 0.30982163548469543,
      "learning_rate": 4.261773746659821e-05,
      "loss": 0.3651,
      "num_input_tokens_seen": 8107144,
      "step": 12440
    },
    {
      "epoch": 6.522536687631027,
      "grad_norm": 0.36160001158714294,
      "learning_rate": 4.260962304515939e-05,
      "loss": 0.4669,
      "num_input_tokens_seen": 8111304,
      "step": 12445
    },
    {
      "epoch": 6.5251572327044025,
      "grad_norm": 0.5517763495445251,
      "learning_rate": 4.2601504939988335e-05,
      "loss": 0.4971,
      "num_input_tokens_seen": 8114056,
      "step": 12450
    },
    {
      "epoch": 6.527777777777778,
      "grad_norm": 0.7994664311408997,
      "learning_rate": 4.259338315278325e-05,
      "loss": 0.5637,
      "num_input_tokens_seen": 8116808,
      "step": 12455
    },
    {
      "epoch": 6.530398322851153,
      "grad_norm": 0.5342909097671509,
      "learning_rate": 4.258525768524313e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 8119496,
      "step": 12460
    },
    {
      "epoch": 6.533018867924528,
      "grad_norm": 0.6564350128173828,
      "learning_rate": 4.2577128539067725e-05,
      "loss": 0.5131,
      "num_input_tokens_seen": 8122728,
      "step": 12465
    },
    {
      "epoch": 6.535639412997903,
      "grad_norm": 0.5309293270111084,
      "learning_rate": 4.256899571595756e-05,
      "loss": 0.587,
      "num_input_tokens_seen": 8126504,
      "step": 12470
    },
    {
      "epoch": 6.538259958071279,
      "grad_norm": 0.6748270392417908,
      "learning_rate": 4.256085921761393e-05,
      "loss": 0.6503,
      "num_input_tokens_seen": 8128776,
      "step": 12475
    },
    {
      "epoch": 6.540880503144654,
      "grad_norm": 0.3559343218803406,
      "learning_rate": 4.25527190457389e-05,
      "loss": 0.5516,
      "num_input_tokens_seen": 8131720,
      "step": 12480
    },
    {
      "epoch": 6.54350104821803,
      "grad_norm": 0.4589006304740906,
      "learning_rate": 4.2544575202035285e-05,
      "loss": 0.5498,
      "num_input_tokens_seen": 8135624,
      "step": 12485
    },
    {
      "epoch": 6.546121593291405,
      "grad_norm": 0.4312724173069,
      "learning_rate": 4.253642768820671e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 8138632,
      "step": 12490
    },
    {
      "epoch": 6.54874213836478,
      "grad_norm": 0.38123834133148193,
      "learning_rate": 4.252827650595752e-05,
      "loss": 0.4771,
      "num_input_tokens_seen": 8142184,
      "step": 12495
    },
    {
      "epoch": 6.551362683438155,
      "grad_norm": 0.3008826971054077,
      "learning_rate": 4.252012165699286e-05,
      "loss": 0.5318,
      "num_input_tokens_seen": 8146792,
      "step": 12500
    },
    {
      "epoch": 6.55398322851153,
      "grad_norm": 0.4829966425895691,
      "learning_rate": 4.2511963143018634e-05,
      "loss": 0.3816,
      "num_input_tokens_seen": 8150504,
      "step": 12505
    },
    {
      "epoch": 6.556603773584905,
      "grad_norm": 0.3618282973766327,
      "learning_rate": 4.25038009657415e-05,
      "loss": 0.4554,
      "num_input_tokens_seen": 8153000,
      "step": 12510
    },
    {
      "epoch": 6.559224318658281,
      "grad_norm": 0.6228624582290649,
      "learning_rate": 4.24956351268689e-05,
      "loss": 0.3838,
      "num_input_tokens_seen": 8155720,
      "step": 12515
    },
    {
      "epoch": 6.561844863731656,
      "grad_norm": 0.4641275107860565,
      "learning_rate": 4.2487465628109036e-05,
      "loss": 0.427,
      "num_input_tokens_seen": 8159400,
      "step": 12520
    },
    {
      "epoch": 6.564465408805032,
      "grad_norm": 0.4795804023742676,
      "learning_rate": 4.247929247117087e-05,
      "loss": 0.3638,
      "num_input_tokens_seen": 8162952,
      "step": 12525
    },
    {
      "epoch": 6.567085953878407,
      "grad_norm": 0.6117480993270874,
      "learning_rate": 4.247111565776413e-05,
      "loss": 0.4664,
      "num_input_tokens_seen": 8166376,
      "step": 12530
    },
    {
      "epoch": 6.569706498951782,
      "grad_norm": 0.7246547937393188,
      "learning_rate": 4.246293518959932e-05,
      "loss": 0.5179,
      "num_input_tokens_seen": 8169064,
      "step": 12535
    },
    {
      "epoch": 6.572327044025157,
      "grad_norm": 0.4864276647567749,
      "learning_rate": 4.245475106838769e-05,
      "loss": 0.5796,
      "num_input_tokens_seen": 8171976,
      "step": 12540
    },
    {
      "epoch": 6.5749475890985325,
      "grad_norm": 0.4459213316440582,
      "learning_rate": 4.2446563295841266e-05,
      "loss": 0.4432,
      "num_input_tokens_seen": 8176200,
      "step": 12545
    },
    {
      "epoch": 6.577568134171908,
      "grad_norm": 0.4772033095359802,
      "learning_rate": 4.243837187367285e-05,
      "loss": 0.4763,
      "num_input_tokens_seen": 8179368,
      "step": 12550
    },
    {
      "epoch": 6.580188679245283,
      "grad_norm": 0.5066973567008972,
      "learning_rate": 4.2430176803595987e-05,
      "loss": 0.3583,
      "num_input_tokens_seen": 8182248,
      "step": 12555
    },
    {
      "epoch": 6.582809224318658,
      "grad_norm": 0.48253247141838074,
      "learning_rate": 4.2421978087324985e-05,
      "loss": 0.58,
      "num_input_tokens_seen": 8184776,
      "step": 12560
    },
    {
      "epoch": 6.585429769392033,
      "grad_norm": 0.3295983076095581,
      "learning_rate": 4.241377572657493e-05,
      "loss": 0.3851,
      "num_input_tokens_seen": 8187976,
      "step": 12565
    },
    {
      "epoch": 6.588050314465409,
      "grad_norm": 0.4904930293560028,
      "learning_rate": 4.2405569723061654e-05,
      "loss": 0.4486,
      "num_input_tokens_seen": 8190984,
      "step": 12570
    },
    {
      "epoch": 6.590670859538784,
      "grad_norm": 0.5544981956481934,
      "learning_rate": 4.239736007850178e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 8194568,
      "step": 12575
    },
    {
      "epoch": 6.59329140461216,
      "grad_norm": 0.5303922295570374,
      "learning_rate": 4.238914679461264e-05,
      "loss": 0.6032,
      "num_input_tokens_seen": 8197512,
      "step": 12580
    },
    {
      "epoch": 6.595911949685535,
      "grad_norm": 0.4297659397125244,
      "learning_rate": 4.2380929873112386e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 8201128,
      "step": 12585
    },
    {
      "epoch": 6.59853249475891,
      "grad_norm": 0.31712618470191956,
      "learning_rate": 4.2372709315719896e-05,
      "loss": 0.3051,
      "num_input_tokens_seen": 8204392,
      "step": 12590
    },
    {
      "epoch": 6.601153039832285,
      "grad_norm": 0.5572468638420105,
      "learning_rate": 4.2364485124154806e-05,
      "loss": 0.4911,
      "num_input_tokens_seen": 8206952,
      "step": 12595
    },
    {
      "epoch": 6.60377358490566,
      "grad_norm": 0.8242070078849792,
      "learning_rate": 4.235625730013754e-05,
      "loss": 0.3943,
      "num_input_tokens_seen": 8209768,
      "step": 12600
    },
    {
      "epoch": 6.606394129979035,
      "grad_norm": 0.4355778396129608,
      "learning_rate": 4.234802584538926e-05,
      "loss": 0.5408,
      "num_input_tokens_seen": 8216680,
      "step": 12605
    },
    {
      "epoch": 6.609014675052411,
      "grad_norm": 0.42061442136764526,
      "learning_rate": 4.2339790761631884e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 8220552,
      "step": 12610
    },
    {
      "epoch": 6.611635220125786,
      "grad_norm": 0.3821529150009155,
      "learning_rate": 4.233155205058811e-05,
      "loss": 0.5016,
      "num_input_tokens_seen": 8225064,
      "step": 12615
    },
    {
      "epoch": 6.614255765199162,
      "grad_norm": 0.46784400939941406,
      "learning_rate": 4.232330971398137e-05,
      "loss": 0.5326,
      "num_input_tokens_seen": 8228168,
      "step": 12620
    },
    {
      "epoch": 6.616876310272537,
      "grad_norm": 1.2460883855819702,
      "learning_rate": 4.2315063753535875e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 8231944,
      "step": 12625
    },
    {
      "epoch": 6.619496855345912,
      "grad_norm": 0.5212637782096863,
      "learning_rate": 4.230681417097658e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 8235848,
      "step": 12630
    },
    {
      "epoch": 6.622117400419287,
      "grad_norm": 0.3154415786266327,
      "learning_rate": 4.229856096802921e-05,
      "loss": 0.5442,
      "num_input_tokens_seen": 8238792,
      "step": 12635
    },
    {
      "epoch": 6.6247379454926625,
      "grad_norm": 0.502610445022583,
      "learning_rate": 4.229030414642024e-05,
      "loss": 0.4386,
      "num_input_tokens_seen": 8242088,
      "step": 12640
    },
    {
      "epoch": 6.627358490566038,
      "grad_norm": 0.4187876582145691,
      "learning_rate": 4.2282043707876896e-05,
      "loss": 0.5801,
      "num_input_tokens_seen": 8244616,
      "step": 12645
    },
    {
      "epoch": 6.629979035639413,
      "grad_norm": 0.5391957759857178,
      "learning_rate": 4.2273779654127175e-05,
      "loss": 0.4108,
      "num_input_tokens_seen": 8247912,
      "step": 12650
    },
    {
      "epoch": 6.632599580712788,
      "grad_norm": 0.497128963470459,
      "learning_rate": 4.226551198689982e-05,
      "loss": 0.466,
      "num_input_tokens_seen": 8251816,
      "step": 12655
    },
    {
      "epoch": 6.635220125786163,
      "grad_norm": 0.4333808124065399,
      "learning_rate": 4.225724070792434e-05,
      "loss": 0.518,
      "num_input_tokens_seen": 8255432,
      "step": 12660
    },
    {
      "epoch": 6.637840670859539,
      "grad_norm": 0.2669420838356018,
      "learning_rate": 4.2248965818930975e-05,
      "loss": 0.3364,
      "num_input_tokens_seen": 8258824,
      "step": 12665
    },
    {
      "epoch": 6.640461215932914,
      "grad_norm": 0.629425048828125,
      "learning_rate": 4.224068732165075e-05,
      "loss": 0.5069,
      "num_input_tokens_seen": 8262888,
      "step": 12670
    },
    {
      "epoch": 6.6430817610062896,
      "grad_norm": 0.528984785079956,
      "learning_rate": 4.223240521781543e-05,
      "loss": 0.3459,
      "num_input_tokens_seen": 8266568,
      "step": 12675
    },
    {
      "epoch": 6.645702306079665,
      "grad_norm": 0.39799684286117554,
      "learning_rate": 4.2224119509157535e-05,
      "loss": 0.5265,
      "num_input_tokens_seen": 8269800,
      "step": 12680
    },
    {
      "epoch": 6.64832285115304,
      "grad_norm": 0.276313453912735,
      "learning_rate": 4.221583019741033e-05,
      "loss": 0.3778,
      "num_input_tokens_seen": 8272936,
      "step": 12685
    },
    {
      "epoch": 6.650943396226415,
      "grad_norm": 0.46954345703125,
      "learning_rate": 4.2207537284307876e-05,
      "loss": 0.4644,
      "num_input_tokens_seen": 8276104,
      "step": 12690
    },
    {
      "epoch": 6.65356394129979,
      "grad_norm": 0.23420120775699615,
      "learning_rate": 4.219924077158493e-05,
      "loss": 0.3728,
      "num_input_tokens_seen": 8278728,
      "step": 12695
    },
    {
      "epoch": 6.656184486373165,
      "grad_norm": 0.47366592288017273,
      "learning_rate": 4.219094066097703e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 8282312,
      "step": 12700
    },
    {
      "epoch": 6.658805031446541,
      "grad_norm": 0.4099594056606293,
      "learning_rate": 4.218263695422047e-05,
      "loss": 0.5042,
      "num_input_tokens_seen": 8285800,
      "step": 12705
    },
    {
      "epoch": 6.661425576519916,
      "grad_norm": 0.4730166792869568,
      "learning_rate": 4.217432965305229e-05,
      "loss": 0.5687,
      "num_input_tokens_seen": 8288648,
      "step": 12710
    },
    {
      "epoch": 6.664046121593291,
      "grad_norm": 0.4878733456134796,
      "learning_rate": 4.2166018759210286e-05,
      "loss": 0.4251,
      "num_input_tokens_seen": 8291688,
      "step": 12715
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 0.6129639744758606,
      "learning_rate": 4.215770427443299e-05,
      "loss": 0.5476,
      "num_input_tokens_seen": 8295400,
      "step": 12720
    },
    {
      "epoch": 6.669287211740042,
      "grad_norm": 0.3302515745162964,
      "learning_rate": 4.214938620045971e-05,
      "loss": 0.4382,
      "num_input_tokens_seen": 8297704,
      "step": 12725
    },
    {
      "epoch": 6.671907756813417,
      "grad_norm": 0.733600914478302,
      "learning_rate": 4.214106453903048e-05,
      "loss": 0.6137,
      "num_input_tokens_seen": 8300808,
      "step": 12730
    },
    {
      "epoch": 6.6745283018867925,
      "grad_norm": 0.4903770387172699,
      "learning_rate": 4.2132739291886114e-05,
      "loss": 0.6433,
      "num_input_tokens_seen": 8304104,
      "step": 12735
    },
    {
      "epoch": 6.677148846960168,
      "grad_norm": 0.5465390682220459,
      "learning_rate": 4.212441046076814e-05,
      "loss": 0.5626,
      "num_input_tokens_seen": 8307784,
      "step": 12740
    },
    {
      "epoch": 6.679769392033543,
      "grad_norm": 0.4152485132217407,
      "learning_rate": 4.211607804741887e-05,
      "loss": 0.4389,
      "num_input_tokens_seen": 8311848,
      "step": 12745
    },
    {
      "epoch": 6.682389937106918,
      "grad_norm": 0.35503360629081726,
      "learning_rate": 4.2107742053581344e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 8315272,
      "step": 12750
    },
    {
      "epoch": 6.685010482180293,
      "grad_norm": 0.5938051342964172,
      "learning_rate": 4.209940248099935e-05,
      "loss": 0.3872,
      "num_input_tokens_seen": 8318376,
      "step": 12755
    },
    {
      "epoch": 6.687631027253669,
      "grad_norm": 0.3832046389579773,
      "learning_rate": 4.2091059331417445e-05,
      "loss": 0.4568,
      "num_input_tokens_seen": 8322248,
      "step": 12760
    },
    {
      "epoch": 6.690251572327044,
      "grad_norm": 0.24846288561820984,
      "learning_rate": 4.208271260658091e-05,
      "loss": 0.3116,
      "num_input_tokens_seen": 8326056,
      "step": 12765
    },
    {
      "epoch": 6.6928721174004195,
      "grad_norm": 0.3894612193107605,
      "learning_rate": 4.2074362308235785e-05,
      "loss": 0.5486,
      "num_input_tokens_seen": 8329608,
      "step": 12770
    },
    {
      "epoch": 6.695492662473795,
      "grad_norm": 0.4100190997123718,
      "learning_rate": 4.206600843812886e-05,
      "loss": 0.4672,
      "num_input_tokens_seen": 8332072,
      "step": 12775
    },
    {
      "epoch": 6.69811320754717,
      "grad_norm": 0.5028148293495178,
      "learning_rate": 4.205765099800767e-05,
      "loss": 0.4711,
      "num_input_tokens_seen": 8334760,
      "step": 12780
    },
    {
      "epoch": 6.700733752620545,
      "grad_norm": 0.3802490830421448,
      "learning_rate": 4.204928998962049e-05,
      "loss": 0.3621,
      "num_input_tokens_seen": 8337736,
      "step": 12785
    },
    {
      "epoch": 6.70335429769392,
      "grad_norm": 0.43799838423728943,
      "learning_rate": 4.204092541471636e-05,
      "loss": 0.3619,
      "num_input_tokens_seen": 8340488,
      "step": 12790
    },
    {
      "epoch": 6.705974842767295,
      "grad_norm": 0.7165044546127319,
      "learning_rate": 4.203255727504504e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 8343176,
      "step": 12795
    },
    {
      "epoch": 6.7085953878406706,
      "grad_norm": 0.39042970538139343,
      "learning_rate": 4.202418557235706e-05,
      "loss": 0.5653,
      "num_input_tokens_seen": 8346376,
      "step": 12800
    },
    {
      "epoch": 6.711215932914046,
      "grad_norm": 0.3094389736652374,
      "learning_rate": 4.201581030840368e-05,
      "loss": 0.4327,
      "num_input_tokens_seen": 8349672,
      "step": 12805
    },
    {
      "epoch": 6.713836477987421,
      "grad_norm": 0.43181532621383667,
      "learning_rate": 4.200743148493691e-05,
      "loss": 0.4902,
      "num_input_tokens_seen": 8352232,
      "step": 12810
    },
    {
      "epoch": 6.716457023060797,
      "grad_norm": 0.29369765520095825,
      "learning_rate": 4.199904910370951e-05,
      "loss": 0.3646,
      "num_input_tokens_seen": 8355592,
      "step": 12815
    },
    {
      "epoch": 6.719077568134172,
      "grad_norm": 0.2968631386756897,
      "learning_rate": 4.1990663166474964e-05,
      "loss": 0.3735,
      "num_input_tokens_seen": 8357736,
      "step": 12820
    },
    {
      "epoch": 6.721698113207547,
      "grad_norm": 0.4301910102367401,
      "learning_rate": 4.1982273674987524e-05,
      "loss": 0.3825,
      "num_input_tokens_seen": 8362568,
      "step": 12825
    },
    {
      "epoch": 6.7243186582809225,
      "grad_norm": 0.3646225035190582,
      "learning_rate": 4.197388063100218e-05,
      "loss": 0.4972,
      "num_input_tokens_seen": 8366152,
      "step": 12830
    },
    {
      "epoch": 6.726939203354298,
      "grad_norm": 1.1419545412063599,
      "learning_rate": 4.196548403627465e-05,
      "loss": 0.4593,
      "num_input_tokens_seen": 8368456,
      "step": 12835
    },
    {
      "epoch": 6.729559748427673,
      "grad_norm": 0.5631918907165527,
      "learning_rate": 4.195708389256142e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 8371560,
      "step": 12840
    },
    {
      "epoch": 6.732180293501048,
      "grad_norm": 0.5623891353607178,
      "learning_rate": 4.194868020161969e-05,
      "loss": 0.4543,
      "num_input_tokens_seen": 8374792,
      "step": 12845
    },
    {
      "epoch": 6.734800838574423,
      "grad_norm": 0.5417872667312622,
      "learning_rate": 4.1940272965207425e-05,
      "loss": 0.4402,
      "num_input_tokens_seen": 8377832,
      "step": 12850
    },
    {
      "epoch": 6.737421383647799,
      "grad_norm": 0.34571292996406555,
      "learning_rate": 4.1931862185083326e-05,
      "loss": 0.5212,
      "num_input_tokens_seen": 8380872,
      "step": 12855
    },
    {
      "epoch": 6.740041928721174,
      "grad_norm": 0.4495999813079834,
      "learning_rate": 4.192344786300682e-05,
      "loss": 0.4484,
      "num_input_tokens_seen": 8384264,
      "step": 12860
    },
    {
      "epoch": 6.7426624737945495,
      "grad_norm": 0.2881941497325897,
      "learning_rate": 4.19150300007381e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 8389064,
      "step": 12865
    },
    {
      "epoch": 6.745283018867925,
      "grad_norm": 0.525759756565094,
      "learning_rate": 4.190660860003806e-05,
      "loss": 0.5382,
      "num_input_tokens_seen": 8391976,
      "step": 12870
    },
    {
      "epoch": 6.7479035639413,
      "grad_norm": 0.3724200129508972,
      "learning_rate": 4.189818366266841e-05,
      "loss": 0.3653,
      "num_input_tokens_seen": 8395176,
      "step": 12875
    },
    {
      "epoch": 6.750524109014675,
      "grad_norm": 0.439109206199646,
      "learning_rate": 4.188975519039151e-05,
      "loss": 0.5359,
      "num_input_tokens_seen": 8398216,
      "step": 12880
    },
    {
      "epoch": 6.75314465408805,
      "grad_norm": 0.31696245074272156,
      "learning_rate": 4.188132318497051e-05,
      "loss": 0.5118,
      "num_input_tokens_seen": 8402120,
      "step": 12885
    },
    {
      "epoch": 6.755765199161425,
      "grad_norm": 0.4854428172111511,
      "learning_rate": 4.1872887648169304e-05,
      "loss": 0.5189,
      "num_input_tokens_seen": 8405736,
      "step": 12890
    },
    {
      "epoch": 6.7583857442348005,
      "grad_norm": 1.4015045166015625,
      "learning_rate": 4.186444858175248e-05,
      "loss": 0.5766,
      "num_input_tokens_seen": 8408808,
      "step": 12895
    },
    {
      "epoch": 6.761006289308176,
      "grad_norm": 0.4540582001209259,
      "learning_rate": 4.1856005987485424e-05,
      "loss": 0.386,
      "num_input_tokens_seen": 8411400,
      "step": 12900
    },
    {
      "epoch": 6.763626834381551,
      "grad_norm": 0.427102655172348,
      "learning_rate": 4.1847559867134215e-05,
      "loss": 0.3354,
      "num_input_tokens_seen": 8414632,
      "step": 12905
    },
    {
      "epoch": 6.766247379454927,
      "grad_norm": 0.7131257653236389,
      "learning_rate": 4.1839110222465693e-05,
      "loss": 0.4866,
      "num_input_tokens_seen": 8417256,
      "step": 12910
    },
    {
      "epoch": 6.768867924528302,
      "grad_norm": 0.41596895456314087,
      "learning_rate": 4.183065705524742e-05,
      "loss": 0.395,
      "num_input_tokens_seen": 8420584,
      "step": 12915
    },
    {
      "epoch": 6.771488469601677,
      "grad_norm": 0.37571799755096436,
      "learning_rate": 4.1822200367247696e-05,
      "loss": 0.5729,
      "num_input_tokens_seen": 8423912,
      "step": 12920
    },
    {
      "epoch": 6.774109014675052,
      "grad_norm": 0.4966253936290741,
      "learning_rate": 4.181374016023559e-05,
      "loss": 0.3561,
      "num_input_tokens_seen": 8427368,
      "step": 12925
    },
    {
      "epoch": 6.776729559748428,
      "grad_norm": 0.3588031828403473,
      "learning_rate": 4.1805276435980854e-05,
      "loss": 0.462,
      "num_input_tokens_seen": 8431176,
      "step": 12930
    },
    {
      "epoch": 6.779350104821803,
      "grad_norm": 0.3412238657474518,
      "learning_rate": 4.1796809196254014e-05,
      "loss": 0.4368,
      "num_input_tokens_seen": 8433768,
      "step": 12935
    },
    {
      "epoch": 6.781970649895178,
      "grad_norm": 0.467033714056015,
      "learning_rate": 4.178833844282632e-05,
      "loss": 0.389,
      "num_input_tokens_seen": 8437384,
      "step": 12940
    },
    {
      "epoch": 6.784591194968553,
      "grad_norm": 0.4445415437221527,
      "learning_rate": 4.177986417746976e-05,
      "loss": 0.3761,
      "num_input_tokens_seen": 8441192,
      "step": 12945
    },
    {
      "epoch": 6.787211740041929,
      "grad_norm": 0.42953094840049744,
      "learning_rate": 4.177138640195704e-05,
      "loss": 0.6035,
      "num_input_tokens_seen": 8446504,
      "step": 12950
    },
    {
      "epoch": 6.789832285115304,
      "grad_norm": 0.4536365270614624,
      "learning_rate": 4.176290511806162e-05,
      "loss": 0.3975,
      "num_input_tokens_seen": 8449384,
      "step": 12955
    },
    {
      "epoch": 6.7924528301886795,
      "grad_norm": 0.581925630569458,
      "learning_rate": 4.1754420327557706e-05,
      "loss": 0.4245,
      "num_input_tokens_seen": 8452456,
      "step": 12960
    },
    {
      "epoch": 6.795073375262055,
      "grad_norm": 0.47923552989959717,
      "learning_rate": 4.1745932032220195e-05,
      "loss": 0.5102,
      "num_input_tokens_seen": 8457000,
      "step": 12965
    },
    {
      "epoch": 6.79769392033543,
      "grad_norm": 0.732830822467804,
      "learning_rate": 4.173744023382474e-05,
      "loss": 0.4465,
      "num_input_tokens_seen": 8461576,
      "step": 12970
    },
    {
      "epoch": 6.800314465408805,
      "grad_norm": 0.6269044280052185,
      "learning_rate": 4.172894493414776e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 8464712,
      "step": 12975
    },
    {
      "epoch": 6.80293501048218,
      "grad_norm": 0.4279073476791382,
      "learning_rate": 4.172044613496634e-05,
      "loss": 0.3767,
      "num_input_tokens_seen": 8467880,
      "step": 12980
    },
    {
      "epoch": 6.805555555555555,
      "grad_norm": 0.43163982033729553,
      "learning_rate": 4.171194383805835e-05,
      "loss": 0.3952,
      "num_input_tokens_seen": 8471208,
      "step": 12985
    },
    {
      "epoch": 6.8081761006289305,
      "grad_norm": 0.3736977279186249,
      "learning_rate": 4.170343804520237e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 8474248,
      "step": 12990
    },
    {
      "epoch": 6.810796645702306,
      "grad_norm": 0.5023439526557922,
      "learning_rate": 4.169492875817771e-05,
      "loss": 0.3867,
      "num_input_tokens_seen": 8477064,
      "step": 12995
    },
    {
      "epoch": 6.813417190775681,
      "grad_norm": 0.6733058094978333,
      "learning_rate": 4.1686415978764415e-05,
      "loss": 0.5095,
      "num_input_tokens_seen": 8479528,
      "step": 13000
    },
    {
      "epoch": 6.816037735849057,
      "grad_norm": 0.506648600101471,
      "learning_rate": 4.167789970874327e-05,
      "loss": 0.3541,
      "num_input_tokens_seen": 8482856,
      "step": 13005
    },
    {
      "epoch": 6.818658280922432,
      "grad_norm": 0.27726462483406067,
      "learning_rate": 4.166937994989578e-05,
      "loss": 0.4706,
      "num_input_tokens_seen": 8486312,
      "step": 13010
    },
    {
      "epoch": 6.821278825995807,
      "grad_norm": 0.42154383659362793,
      "learning_rate": 4.166085670400418e-05,
      "loss": 0.513,
      "num_input_tokens_seen": 8489096,
      "step": 13015
    },
    {
      "epoch": 6.823899371069182,
      "grad_norm": 0.24860668182373047,
      "learning_rate": 4.165232997285143e-05,
      "loss": 0.5108,
      "num_input_tokens_seen": 8493448,
      "step": 13020
    },
    {
      "epoch": 6.826519916142558,
      "grad_norm": 0.35221025347709656,
      "learning_rate": 4.164379975822124e-05,
      "loss": 0.3403,
      "num_input_tokens_seen": 8496968,
      "step": 13025
    },
    {
      "epoch": 6.829140461215933,
      "grad_norm": 1.5670841932296753,
      "learning_rate": 4.1635266061898015e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 8500616,
      "step": 13030
    },
    {
      "epoch": 6.831761006289308,
      "grad_norm": 0.4190632998943329,
      "learning_rate": 4.162672888566691e-05,
      "loss": 0.4897,
      "num_input_tokens_seen": 8504136,
      "step": 13035
    },
    {
      "epoch": 6.834381551362683,
      "grad_norm": 0.42344680428504944,
      "learning_rate": 4.1618188231313816e-05,
      "loss": 0.5358,
      "num_input_tokens_seen": 8507464,
      "step": 13040
    },
    {
      "epoch": 6.837002096436059,
      "grad_norm": 0.6306162476539612,
      "learning_rate": 4.1609644100625335e-05,
      "loss": 0.533,
      "num_input_tokens_seen": 8510216,
      "step": 13045
    },
    {
      "epoch": 6.839622641509434,
      "grad_norm": 0.4199705123901367,
      "learning_rate": 4.16010964953888e-05,
      "loss": 0.5183,
      "num_input_tokens_seen": 8513960,
      "step": 13050
    },
    {
      "epoch": 6.8422431865828095,
      "grad_norm": 0.3047659695148468,
      "learning_rate": 4.159254541739227e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 8517768,
      "step": 13055
    },
    {
      "epoch": 6.844863731656185,
      "grad_norm": 0.32190361618995667,
      "learning_rate": 4.158399086842454e-05,
      "loss": 0.5479,
      "num_input_tokens_seen": 8521640,
      "step": 13060
    },
    {
      "epoch": 6.84748427672956,
      "grad_norm": 0.4986812174320221,
      "learning_rate": 4.157543285027512e-05,
      "loss": 0.4228,
      "num_input_tokens_seen": 8524968,
      "step": 13065
    },
    {
      "epoch": 6.850104821802935,
      "grad_norm": 0.7325087189674377,
      "learning_rate": 4.156687136473425e-05,
      "loss": 0.6613,
      "num_input_tokens_seen": 8528200,
      "step": 13070
    },
    {
      "epoch": 6.85272536687631,
      "grad_norm": 0.4300149083137512,
      "learning_rate": 4.1558306413592894e-05,
      "loss": 0.4617,
      "num_input_tokens_seen": 8531144,
      "step": 13075
    },
    {
      "epoch": 6.855345911949685,
      "grad_norm": 0.4755471646785736,
      "learning_rate": 4.1549737998642745e-05,
      "loss": 0.4889,
      "num_input_tokens_seen": 8535240,
      "step": 13080
    },
    {
      "epoch": 6.8579664570230605,
      "grad_norm": 0.31109118461608887,
      "learning_rate": 4.154116612167622e-05,
      "loss": 0.5159,
      "num_input_tokens_seen": 8537352,
      "step": 13085
    },
    {
      "epoch": 6.860587002096436,
      "grad_norm": 0.4367888569831848,
      "learning_rate": 4.1532590784486444e-05,
      "loss": 0.4077,
      "num_input_tokens_seen": 8540776,
      "step": 13090
    },
    {
      "epoch": 6.863207547169811,
      "grad_norm": 0.520558774471283,
      "learning_rate": 4.1524011988867294e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 8543944,
      "step": 13095
    },
    {
      "epoch": 6.865828092243187,
      "grad_norm": 0.5407935380935669,
      "learning_rate": 4.151542973661336e-05,
      "loss": 0.473,
      "num_input_tokens_seen": 8546984,
      "step": 13100
    },
    {
      "epoch": 6.868448637316562,
      "grad_norm": 0.8442209959030151,
      "learning_rate": 4.150684402951994e-05,
      "loss": 0.8054,
      "num_input_tokens_seen": 8550376,
      "step": 13105
    },
    {
      "epoch": 6.871069182389937,
      "grad_norm": 0.6240577101707458,
      "learning_rate": 4.149825486938306e-05,
      "loss": 0.4991,
      "num_input_tokens_seen": 8553512,
      "step": 13110
    },
    {
      "epoch": 6.873689727463312,
      "grad_norm": 0.3810657560825348,
      "learning_rate": 4.148966225799949e-05,
      "loss": 0.4372,
      "num_input_tokens_seen": 8557320,
      "step": 13115
    },
    {
      "epoch": 6.876310272536688,
      "grad_norm": 0.43853455781936646,
      "learning_rate": 4.14810661971667e-05,
      "loss": 0.4106,
      "num_input_tokens_seen": 8559848,
      "step": 13120
    },
    {
      "epoch": 6.878930817610063,
      "grad_norm": 0.25511494278907776,
      "learning_rate": 4.147246668868289e-05,
      "loss": 0.307,
      "num_input_tokens_seen": 8563400,
      "step": 13125
    },
    {
      "epoch": 6.881551362683438,
      "grad_norm": 0.3862043619155884,
      "learning_rate": 4.1463863734346964e-05,
      "loss": 0.4461,
      "num_input_tokens_seen": 8566728,
      "step": 13130
    },
    {
      "epoch": 6.884171907756813,
      "grad_norm": 0.6016697883605957,
      "learning_rate": 4.145525733595859e-05,
      "loss": 0.4298,
      "num_input_tokens_seen": 8570056,
      "step": 13135
    },
    {
      "epoch": 6.886792452830189,
      "grad_norm": 0.45513954758644104,
      "learning_rate": 4.14466474953181e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 8573064,
      "step": 13140
    },
    {
      "epoch": 6.889412997903564,
      "grad_norm": 0.41762733459472656,
      "learning_rate": 4.14380342142266e-05,
      "loss": 0.5695,
      "num_input_tokens_seen": 8577640,
      "step": 13145
    },
    {
      "epoch": 6.8920335429769395,
      "grad_norm": 0.3936237096786499,
      "learning_rate": 4.142941749448587e-05,
      "loss": 0.5819,
      "num_input_tokens_seen": 8581256,
      "step": 13150
    },
    {
      "epoch": 6.894654088050315,
      "grad_norm": 0.3691321015357971,
      "learning_rate": 4.142079733789843e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 8584168,
      "step": 13155
    },
    {
      "epoch": 6.89727463312369,
      "grad_norm": 0.3115008473396301,
      "learning_rate": 4.141217374626753e-05,
      "loss": 0.4813,
      "num_input_tokens_seen": 8587432,
      "step": 13160
    },
    {
      "epoch": 6.899895178197065,
      "grad_norm": 0.3568492829799652,
      "learning_rate": 4.140354672139713e-05,
      "loss": 0.4952,
      "num_input_tokens_seen": 8590728,
      "step": 13165
    },
    {
      "epoch": 6.90251572327044,
      "grad_norm": 0.42564159631729126,
      "learning_rate": 4.1394916265091886e-05,
      "loss": 0.3879,
      "num_input_tokens_seen": 8594120,
      "step": 13170
    },
    {
      "epoch": 6.905136268343815,
      "grad_norm": 0.3500388562679291,
      "learning_rate": 4.1386282379157214e-05,
      "loss": 0.4788,
      "num_input_tokens_seen": 8596904,
      "step": 13175
    },
    {
      "epoch": 6.9077568134171905,
      "grad_norm": 0.3972729742527008,
      "learning_rate": 4.137764506539921e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 8599848,
      "step": 13180
    },
    {
      "epoch": 6.910377358490566,
      "grad_norm": 0.28626060485839844,
      "learning_rate": 4.136900432562471e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 8603272,
      "step": 13185
    },
    {
      "epoch": 6.912997903563941,
      "grad_norm": 0.3461458086967468,
      "learning_rate": 4.136036016164123e-05,
      "loss": 0.4475,
      "num_input_tokens_seen": 8607240,
      "step": 13190
    },
    {
      "epoch": 6.915618448637317,
      "grad_norm": 1.2022186517715454,
      "learning_rate": 4.135171257525707e-05,
      "loss": 0.6132,
      "num_input_tokens_seen": 8613160,
      "step": 13195
    },
    {
      "epoch": 6.918238993710692,
      "grad_norm": 0.4015505015850067,
      "learning_rate": 4.1343061568281195e-05,
      "loss": 0.4868,
      "num_input_tokens_seen": 8617480,
      "step": 13200
    },
    {
      "epoch": 6.920859538784067,
      "grad_norm": 0.33089104294776917,
      "learning_rate": 4.133440714252328e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 8620520,
      "step": 13205
    },
    {
      "epoch": 6.923480083857442,
      "grad_norm": 0.44646766781806946,
      "learning_rate": 4.132574929979375e-05,
      "loss": 0.3305,
      "num_input_tokens_seen": 8626568,
      "step": 13210
    },
    {
      "epoch": 6.926100628930818,
      "grad_norm": 0.34484246373176575,
      "learning_rate": 4.1317088041903714e-05,
      "loss": 0.4912,
      "num_input_tokens_seen": 8630184,
      "step": 13215
    },
    {
      "epoch": 6.928721174004193,
      "grad_norm": 0.508277416229248,
      "learning_rate": 4.1308423370665015e-05,
      "loss": 0.4297,
      "num_input_tokens_seen": 8633576,
      "step": 13220
    },
    {
      "epoch": 6.931341719077568,
      "grad_norm": 0.3603527545928955,
      "learning_rate": 4.1299755287890206e-05,
      "loss": 0.4568,
      "num_input_tokens_seen": 8636520,
      "step": 13225
    },
    {
      "epoch": 6.933962264150943,
      "grad_norm": 0.429415225982666,
      "learning_rate": 4.129108379539255e-05,
      "loss": 0.3376,
      "num_input_tokens_seen": 8639816,
      "step": 13230
    },
    {
      "epoch": 6.936582809224319,
      "grad_norm": 0.5516642332077026,
      "learning_rate": 4.128240889498602e-05,
      "loss": 0.5149,
      "num_input_tokens_seen": 8642888,
      "step": 13235
    },
    {
      "epoch": 6.939203354297694,
      "grad_norm": 0.42333686351776123,
      "learning_rate": 4.127373058848531e-05,
      "loss": 0.4855,
      "num_input_tokens_seen": 8645704,
      "step": 13240
    },
    {
      "epoch": 6.9418238993710695,
      "grad_norm": 0.39092156291007996,
      "learning_rate": 4.126504887770582e-05,
      "loss": 0.4163,
      "num_input_tokens_seen": 8648648,
      "step": 13245
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 0.4545060098171234,
      "learning_rate": 4.1256363764463666e-05,
      "loss": 0.4102,
      "num_input_tokens_seen": 8651560,
      "step": 13250
    },
    {
      "epoch": 6.94706498951782,
      "grad_norm": 0.2926727533340454,
      "learning_rate": 4.1247675250575676e-05,
      "loss": 0.5441,
      "num_input_tokens_seen": 8655208,
      "step": 13255
    },
    {
      "epoch": 6.949685534591195,
      "grad_norm": 0.4183071255683899,
      "learning_rate": 4.1238983337859396e-05,
      "loss": 0.3259,
      "num_input_tokens_seen": 8657576,
      "step": 13260
    },
    {
      "epoch": 6.95230607966457,
      "grad_norm": 0.6489309668540955,
      "learning_rate": 4.123028802813305e-05,
      "loss": 0.705,
      "num_input_tokens_seen": 8660872,
      "step": 13265
    },
    {
      "epoch": 6.954926624737945,
      "grad_norm": 0.4243435859680176,
      "learning_rate": 4.1221589323215625e-05,
      "loss": 0.4541,
      "num_input_tokens_seen": 8665192,
      "step": 13270
    },
    {
      "epoch": 6.9575471698113205,
      "grad_norm": 0.1931595802307129,
      "learning_rate": 4.121288722492678e-05,
      "loss": 0.4356,
      "num_input_tokens_seen": 8670088,
      "step": 13275
    },
    {
      "epoch": 6.960167714884696,
      "grad_norm": 0.8305836915969849,
      "learning_rate": 4.120418173508688e-05,
      "loss": 0.4797,
      "num_input_tokens_seen": 8673576,
      "step": 13280
    },
    {
      "epoch": 6.962788259958071,
      "grad_norm": 0.4343101978302002,
      "learning_rate": 4.1195472855517045e-05,
      "loss": 0.6538,
      "num_input_tokens_seen": 8676424,
      "step": 13285
    },
    {
      "epoch": 6.965408805031447,
      "grad_norm": 0.3996300995349884,
      "learning_rate": 4.1186760588039045e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 8679912,
      "step": 13290
    },
    {
      "epoch": 6.968029350104822,
      "grad_norm": 0.36325353384017944,
      "learning_rate": 4.11780449344754e-05,
      "loss": 0.7038,
      "num_input_tokens_seen": 8683880,
      "step": 13295
    },
    {
      "epoch": 6.970649895178197,
      "grad_norm": 0.364572674036026,
      "learning_rate": 4.1169325896649326e-05,
      "loss": 0.5436,
      "num_input_tokens_seen": 8687784,
      "step": 13300
    },
    {
      "epoch": 6.973270440251572,
      "grad_norm": 0.3690241575241089,
      "learning_rate": 4.116060347638474e-05,
      "loss": 0.4329,
      "num_input_tokens_seen": 8690920,
      "step": 13305
    },
    {
      "epoch": 6.975890985324948,
      "grad_norm": 0.7728114724159241,
      "learning_rate": 4.115187767550628e-05,
      "loss": 0.3641,
      "num_input_tokens_seen": 8693960,
      "step": 13310
    },
    {
      "epoch": 6.978511530398323,
      "grad_norm": 0.25941571593284607,
      "learning_rate": 4.114314849583928e-05,
      "loss": 0.5165,
      "num_input_tokens_seen": 8697640,
      "step": 13315
    },
    {
      "epoch": 6.981132075471698,
      "grad_norm": 0.3901427388191223,
      "learning_rate": 4.113441593920979e-05,
      "loss": 0.6826,
      "num_input_tokens_seen": 8699912,
      "step": 13320
    },
    {
      "epoch": 6.983752620545073,
      "grad_norm": 0.4829004108905792,
      "learning_rate": 4.112568000744454e-05,
      "loss": 0.6071,
      "num_input_tokens_seen": 8703240,
      "step": 13325
    },
    {
      "epoch": 6.986373165618449,
      "grad_norm": 0.43758806586265564,
      "learning_rate": 4.1116940702371025e-05,
      "loss": 0.3752,
      "num_input_tokens_seen": 8706376,
      "step": 13330
    },
    {
      "epoch": 6.988993710691824,
      "grad_norm": 0.5529744029045105,
      "learning_rate": 4.1108198025817376e-05,
      "loss": 0.5813,
      "num_input_tokens_seen": 8709224,
      "step": 13335
    },
    {
      "epoch": 6.9916142557651995,
      "grad_norm": 0.9227559566497803,
      "learning_rate": 4.1099451979612476e-05,
      "loss": 0.5445,
      "num_input_tokens_seen": 8711816,
      "step": 13340
    },
    {
      "epoch": 6.994234800838575,
      "grad_norm": 0.746094286441803,
      "learning_rate": 4.109070256558589e-05,
      "loss": 0.4638,
      "num_input_tokens_seen": 8715496,
      "step": 13345
    },
    {
      "epoch": 6.99685534591195,
      "grad_norm": 0.4010784924030304,
      "learning_rate": 4.108194978556791e-05,
      "loss": 0.3928,
      "num_input_tokens_seen": 8718312,
      "step": 13350
    },
    {
      "epoch": 6.999475890985325,
      "grad_norm": 0.5543643236160278,
      "learning_rate": 4.1073193641389495e-05,
      "loss": 0.4485,
      "num_input_tokens_seen": 8722632,
      "step": 13355
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.4815897047519684,
      "eval_runtime": 9.7877,
      "eval_samples_per_second": 86.639,
      "eval_steps_per_second": 21.66,
      "num_input_tokens_seen": 8722744,
      "step": 13356
    },
    {
      "epoch": 7.0020964360587,
      "grad_norm": 0.5279285311698914,
      "learning_rate": 4.1064434134882345e-05,
      "loss": 0.4613,
      "num_input_tokens_seen": 8725944,
      "step": 13360
    },
    {
      "epoch": 7.004716981132075,
      "grad_norm": 0.36736565828323364,
      "learning_rate": 4.1055671267878847e-05,
      "loss": 0.4701,
      "num_input_tokens_seen": 8729592,
      "step": 13365
    },
    {
      "epoch": 7.0073375262054505,
      "grad_norm": 0.5410627722740173,
      "learning_rate": 4.1046905042212106e-05,
      "loss": 0.4097,
      "num_input_tokens_seen": 8734264,
      "step": 13370
    },
    {
      "epoch": 7.009958071278826,
      "grad_norm": 0.29611679911613464,
      "learning_rate": 4.103813545971589e-05,
      "loss": 0.4008,
      "num_input_tokens_seen": 8738744,
      "step": 13375
    },
    {
      "epoch": 7.012578616352202,
      "grad_norm": 0.44453951716423035,
      "learning_rate": 4.10293625222247e-05,
      "loss": 0.3924,
      "num_input_tokens_seen": 8741720,
      "step": 13380
    },
    {
      "epoch": 7.015199161425577,
      "grad_norm": 0.3929100036621094,
      "learning_rate": 4.102058623157376e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 8745112,
      "step": 13385
    },
    {
      "epoch": 7.017819706498952,
      "grad_norm": 1.352988362312317,
      "learning_rate": 4.101180658959894e-05,
      "loss": 0.4212,
      "num_input_tokens_seen": 8747960,
      "step": 13390
    },
    {
      "epoch": 7.020440251572327,
      "grad_norm": 0.6475703120231628,
      "learning_rate": 4.1003023598136866e-05,
      "loss": 0.4951,
      "num_input_tokens_seen": 8750808,
      "step": 13395
    },
    {
      "epoch": 7.023060796645702,
      "grad_norm": 0.39619287848472595,
      "learning_rate": 4.099423725902482e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 8753368,
      "step": 13400
    },
    {
      "epoch": 7.0256813417190775,
      "grad_norm": 0.7500343918800354,
      "learning_rate": 4.098544757410081e-05,
      "loss": 0.5035,
      "num_input_tokens_seen": 8756056,
      "step": 13405
    },
    {
      "epoch": 7.028301886792453,
      "grad_norm": 0.6658151149749756,
      "learning_rate": 4.097665454520354e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 8759256,
      "step": 13410
    },
    {
      "epoch": 7.030922431865828,
      "grad_norm": 0.7124875783920288,
      "learning_rate": 4.096785817417241e-05,
      "loss": 0.3941,
      "num_input_tokens_seen": 8763032,
      "step": 13415
    },
    {
      "epoch": 7.033542976939203,
      "grad_norm": 0.8371536731719971,
      "learning_rate": 4.0959058462847524e-05,
      "loss": 0.3383,
      "num_input_tokens_seen": 8766072,
      "step": 13420
    },
    {
      "epoch": 7.036163522012578,
      "grad_norm": 0.6487521529197693,
      "learning_rate": 4.095025541306968e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 8769976,
      "step": 13425
    },
    {
      "epoch": 7.038784067085954,
      "grad_norm": 0.38031625747680664,
      "learning_rate": 4.0941449026680374e-05,
      "loss": 0.5222,
      "num_input_tokens_seen": 8773528,
      "step": 13430
    },
    {
      "epoch": 7.0414046121593294,
      "grad_norm": 0.8188082575798035,
      "learning_rate": 4.09326393055218e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 8776440,
      "step": 13435
    },
    {
      "epoch": 7.044025157232705,
      "grad_norm": 0.32668742537498474,
      "learning_rate": 4.0923826251436845e-05,
      "loss": 0.3411,
      "num_input_tokens_seen": 8780408,
      "step": 13440
    },
    {
      "epoch": 7.04664570230608,
      "grad_norm": 0.3641597032546997,
      "learning_rate": 4.0915009866269115e-05,
      "loss": 0.4074,
      "num_input_tokens_seen": 8782872,
      "step": 13445
    },
    {
      "epoch": 7.049266247379455,
      "grad_norm": 0.2689879834651947,
      "learning_rate": 4.090619015186288e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 8786776,
      "step": 13450
    },
    {
      "epoch": 7.05188679245283,
      "grad_norm": 0.39620816707611084,
      "learning_rate": 4.0897367110063145e-05,
      "loss": 0.502,
      "num_input_tokens_seen": 8791576,
      "step": 13455
    },
    {
      "epoch": 7.054507337526205,
      "grad_norm": 0.5707810521125793,
      "learning_rate": 4.088854074271557e-05,
      "loss": 0.424,
      "num_input_tokens_seen": 8794072,
      "step": 13460
    },
    {
      "epoch": 7.0571278825995805,
      "grad_norm": 0.3120073974132538,
      "learning_rate": 4.0879711051666534e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 8798424,
      "step": 13465
    },
    {
      "epoch": 7.059748427672956,
      "grad_norm": 0.5950521230697632,
      "learning_rate": 4.087087803876313e-05,
      "loss": 0.3651,
      "num_input_tokens_seen": 8801208,
      "step": 13470
    },
    {
      "epoch": 7.062368972746331,
      "grad_norm": 0.4643019139766693,
      "learning_rate": 4.086204170585308e-05,
      "loss": 0.4868,
      "num_input_tokens_seen": 8804472,
      "step": 13475
    },
    {
      "epoch": 7.064989517819707,
      "grad_norm": 0.30938225984573364,
      "learning_rate": 4.085320205478488e-05,
      "loss": 0.5134,
      "num_input_tokens_seen": 8808152,
      "step": 13480
    },
    {
      "epoch": 7.067610062893082,
      "grad_norm": 0.35379907488822937,
      "learning_rate": 4.084435908740768e-05,
      "loss": 0.4137,
      "num_input_tokens_seen": 8811352,
      "step": 13485
    },
    {
      "epoch": 7.070230607966457,
      "grad_norm": 0.967657744884491,
      "learning_rate": 4.0835512805571306e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 8814328,
      "step": 13490
    },
    {
      "epoch": 7.072851153039832,
      "grad_norm": 0.2475297451019287,
      "learning_rate": 4.082666321112633e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 8817976,
      "step": 13495
    },
    {
      "epoch": 7.0754716981132075,
      "grad_norm": 0.36403200030326843,
      "learning_rate": 4.081781030592396e-05,
      "loss": 0.3757,
      "num_input_tokens_seen": 8820856,
      "step": 13500
    },
    {
      "epoch": 7.078092243186583,
      "grad_norm": 0.5986400246620178,
      "learning_rate": 4.080895409181613e-05,
      "loss": 0.5557,
      "num_input_tokens_seen": 8824120,
      "step": 13505
    },
    {
      "epoch": 7.080712788259958,
      "grad_norm": 0.4531683325767517,
      "learning_rate": 4.080009457065547e-05,
      "loss": 0.4874,
      "num_input_tokens_seen": 8827192,
      "step": 13510
    },
    {
      "epoch": 7.083333333333333,
      "grad_norm": 0.5725570917129517,
      "learning_rate": 4.0791231744295286e-05,
      "loss": 0.3874,
      "num_input_tokens_seen": 8831064,
      "step": 13515
    },
    {
      "epoch": 7.085953878406708,
      "grad_norm": 0.29229670763015747,
      "learning_rate": 4.078236561458957e-05,
      "loss": 0.5392,
      "num_input_tokens_seen": 8834200,
      "step": 13520
    },
    {
      "epoch": 7.088574423480084,
      "grad_norm": 0.532781183719635,
      "learning_rate": 4.077349618339302e-05,
      "loss": 0.3758,
      "num_input_tokens_seen": 8837240,
      "step": 13525
    },
    {
      "epoch": 7.091194968553459,
      "grad_norm": 0.31055304408073425,
      "learning_rate": 4.076462345256103e-05,
      "loss": 0.6111,
      "num_input_tokens_seen": 8841976,
      "step": 13530
    },
    {
      "epoch": 7.093815513626835,
      "grad_norm": 0.5477327704429626,
      "learning_rate": 4.0755747423949655e-05,
      "loss": 0.3757,
      "num_input_tokens_seen": 8845336,
      "step": 13535
    },
    {
      "epoch": 7.09643605870021,
      "grad_norm": 0.4068683981895447,
      "learning_rate": 4.074686809941568e-05,
      "loss": 0.429,
      "num_input_tokens_seen": 8848184,
      "step": 13540
    },
    {
      "epoch": 7.099056603773585,
      "grad_norm": 1.2393945455551147,
      "learning_rate": 4.0737985480816546e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 8850840,
      "step": 13545
    },
    {
      "epoch": 7.10167714884696,
      "grad_norm": 0.8993611931800842,
      "learning_rate": 4.0729099570010395e-05,
      "loss": 0.573,
      "num_input_tokens_seen": 8854296,
      "step": 13550
    },
    {
      "epoch": 7.104297693920335,
      "grad_norm": 0.39653539657592773,
      "learning_rate": 4.072021036885607e-05,
      "loss": 0.418,
      "num_input_tokens_seen": 8857560,
      "step": 13555
    },
    {
      "epoch": 7.1069182389937104,
      "grad_norm": 0.24457456171512604,
      "learning_rate": 4.0711317879213075e-05,
      "loss": 0.4296,
      "num_input_tokens_seen": 8861496,
      "step": 13560
    },
    {
      "epoch": 7.109538784067086,
      "grad_norm": 0.4847281575202942,
      "learning_rate": 4.070242210294162e-05,
      "loss": 0.4574,
      "num_input_tokens_seen": 8864344,
      "step": 13565
    },
    {
      "epoch": 7.112159329140461,
      "grad_norm": 0.2521923780441284,
      "learning_rate": 4.069352304190261e-05,
      "loss": 0.5333,
      "num_input_tokens_seen": 8868152,
      "step": 13570
    },
    {
      "epoch": 7.114779874213837,
      "grad_norm": 0.5502330660820007,
      "learning_rate": 4.068462069795761e-05,
      "loss": 0.3279,
      "num_input_tokens_seen": 8870776,
      "step": 13575
    },
    {
      "epoch": 7.117400419287212,
      "grad_norm": 0.6598685383796692,
      "learning_rate": 4.067571507296891e-05,
      "loss": 0.5478,
      "num_input_tokens_seen": 8873944,
      "step": 13580
    },
    {
      "epoch": 7.120020964360587,
      "grad_norm": 0.8830129504203796,
      "learning_rate": 4.0666806168799455e-05,
      "loss": 0.4906,
      "num_input_tokens_seen": 8876792,
      "step": 13585
    },
    {
      "epoch": 7.122641509433962,
      "grad_norm": 0.8174329996109009,
      "learning_rate": 4.0657893987312876e-05,
      "loss": 0.5525,
      "num_input_tokens_seen": 8879672,
      "step": 13590
    },
    {
      "epoch": 7.1252620545073375,
      "grad_norm": 0.7140295505523682,
      "learning_rate": 4.064897853037352e-05,
      "loss": 0.3538,
      "num_input_tokens_seen": 8882904,
      "step": 13595
    },
    {
      "epoch": 7.127882599580713,
      "grad_norm": 0.565968930721283,
      "learning_rate": 4.064005979984639e-05,
      "loss": 0.3684,
      "num_input_tokens_seen": 8886328,
      "step": 13600
    },
    {
      "epoch": 7.130503144654088,
      "grad_norm": 0.8076925873756409,
      "learning_rate": 4.0631137797597176e-05,
      "loss": 0.4398,
      "num_input_tokens_seen": 8888888,
      "step": 13605
    },
    {
      "epoch": 7.133123689727463,
      "grad_norm": 0.4279315173625946,
      "learning_rate": 4.0622212525492266e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 8892536,
      "step": 13610
    },
    {
      "epoch": 7.135744234800838,
      "grad_norm": 0.3000747859477997,
      "learning_rate": 4.061328398539872e-05,
      "loss": 0.4218,
      "num_input_tokens_seen": 8895768,
      "step": 13615
    },
    {
      "epoch": 7.138364779874214,
      "grad_norm": 0.4650534391403198,
      "learning_rate": 4.060435217918429e-05,
      "loss": 0.4414,
      "num_input_tokens_seen": 8899320,
      "step": 13620
    },
    {
      "epoch": 7.140985324947589,
      "grad_norm": 0.5654129981994629,
      "learning_rate": 4.0595417108717405e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 8901816,
      "step": 13625
    },
    {
      "epoch": 7.143605870020965,
      "grad_norm": 0.44210368394851685,
      "learning_rate": 4.0586478775867185e-05,
      "loss": 0.3707,
      "num_input_tokens_seen": 8904984,
      "step": 13630
    },
    {
      "epoch": 7.14622641509434,
      "grad_norm": 0.35072243213653564,
      "learning_rate": 4.057753718250343e-05,
      "loss": 0.4255,
      "num_input_tokens_seen": 8909272,
      "step": 13635
    },
    {
      "epoch": 7.148846960167715,
      "grad_norm": 0.3204178512096405,
      "learning_rate": 4.056859233049661e-05,
      "loss": 0.507,
      "num_input_tokens_seen": 8912088,
      "step": 13640
    },
    {
      "epoch": 7.15146750524109,
      "grad_norm": 0.4108503758907318,
      "learning_rate": 4.05596442217179e-05,
      "loss": 0.5251,
      "num_input_tokens_seen": 8915160,
      "step": 13645
    },
    {
      "epoch": 7.154088050314465,
      "grad_norm": 0.2558586895465851,
      "learning_rate": 4.055069285803912e-05,
      "loss": 0.404,
      "num_input_tokens_seen": 8918200,
      "step": 13650
    },
    {
      "epoch": 7.15670859538784,
      "grad_norm": 0.37067949771881104,
      "learning_rate": 4.0541738241332815e-05,
      "loss": 0.3567,
      "num_input_tokens_seen": 8921464,
      "step": 13655
    },
    {
      "epoch": 7.159329140461216,
      "grad_norm": 0.5707855820655823,
      "learning_rate": 4.053278037347217e-05,
      "loss": 0.4613,
      "num_input_tokens_seen": 8924344,
      "step": 13660
    },
    {
      "epoch": 7.161949685534591,
      "grad_norm": 0.6186684370040894,
      "learning_rate": 4.052381925633109e-05,
      "loss": 0.4273,
      "num_input_tokens_seen": 8927352,
      "step": 13665
    },
    {
      "epoch": 7.164570230607967,
      "grad_norm": 0.4373953342437744,
      "learning_rate": 4.051485489178412e-05,
      "loss": 0.512,
      "num_input_tokens_seen": 8930936,
      "step": 13670
    },
    {
      "epoch": 7.167190775681342,
      "grad_norm": 0.8721173405647278,
      "learning_rate": 4.05058872817065e-05,
      "loss": 0.5333,
      "num_input_tokens_seen": 8934296,
      "step": 13675
    },
    {
      "epoch": 7.169811320754717,
      "grad_norm": 0.28067252039909363,
      "learning_rate": 4.0496916427974177e-05,
      "loss": 0.3376,
      "num_input_tokens_seen": 8937016,
      "step": 13680
    },
    {
      "epoch": 7.172431865828092,
      "grad_norm": 0.41329821944236755,
      "learning_rate": 4.048794233246373e-05,
      "loss": 0.3759,
      "num_input_tokens_seen": 8940792,
      "step": 13685
    },
    {
      "epoch": 7.1750524109014675,
      "grad_norm": 0.3997954726219177,
      "learning_rate": 4.0478964997052435e-05,
      "loss": 0.5115,
      "num_input_tokens_seen": 8944024,
      "step": 13690
    },
    {
      "epoch": 7.177672955974843,
      "grad_norm": 0.5987617373466492,
      "learning_rate": 4.046998442361827e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 8946872,
      "step": 13695
    },
    {
      "epoch": 7.180293501048218,
      "grad_norm": 0.32670849561691284,
      "learning_rate": 4.046100061403984e-05,
      "loss": 0.4333,
      "num_input_tokens_seen": 8950168,
      "step": 13700
    },
    {
      "epoch": 7.182914046121593,
      "grad_norm": 0.8168419003486633,
      "learning_rate": 4.045201357019648e-05,
      "loss": 0.4491,
      "num_input_tokens_seen": 8953112,
      "step": 13705
    },
    {
      "epoch": 7.185534591194968,
      "grad_norm": 0.6483847498893738,
      "learning_rate": 4.044302329396816e-05,
      "loss": 0.3801,
      "num_input_tokens_seen": 8956312,
      "step": 13710
    },
    {
      "epoch": 7.188155136268344,
      "grad_norm": 0.6984851956367493,
      "learning_rate": 4.0434029787235564e-05,
      "loss": 0.4078,
      "num_input_tokens_seen": 8959032,
      "step": 13715
    },
    {
      "epoch": 7.190775681341719,
      "grad_norm": 0.6688817143440247,
      "learning_rate": 4.042503305188e-05,
      "loss": 0.5697,
      "num_input_tokens_seen": 8962744,
      "step": 13720
    },
    {
      "epoch": 7.193396226415095,
      "grad_norm": 0.37428390979766846,
      "learning_rate": 4.041603308978351e-05,
      "loss": 0.4759,
      "num_input_tokens_seen": 8966520,
      "step": 13725
    },
    {
      "epoch": 7.19601677148847,
      "grad_norm": 0.44243958592414856,
      "learning_rate": 4.040702990282878e-05,
      "loss": 0.5264,
      "num_input_tokens_seen": 8969304,
      "step": 13730
    },
    {
      "epoch": 7.198637316561845,
      "grad_norm": 0.27299854159355164,
      "learning_rate": 4.039802349289914e-05,
      "loss": 0.5155,
      "num_input_tokens_seen": 8973432,
      "step": 13735
    },
    {
      "epoch": 7.20125786163522,
      "grad_norm": 0.427293062210083,
      "learning_rate": 4.038901386187868e-05,
      "loss": 0.4965,
      "num_input_tokens_seen": 8976248,
      "step": 13740
    },
    {
      "epoch": 7.203878406708595,
      "grad_norm": 0.38506579399108887,
      "learning_rate": 4.038000101165208e-05,
      "loss": 0.451,
      "num_input_tokens_seen": 8980792,
      "step": 13745
    },
    {
      "epoch": 7.20649895178197,
      "grad_norm": 0.4703739583492279,
      "learning_rate": 4.0370984944104725e-05,
      "loss": 0.5576,
      "num_input_tokens_seen": 8983480,
      "step": 13750
    },
    {
      "epoch": 7.209119496855346,
      "grad_norm": 0.8528887033462524,
      "learning_rate": 4.0361965661122695e-05,
      "loss": 0.4848,
      "num_input_tokens_seen": 8986200,
      "step": 13755
    },
    {
      "epoch": 7.211740041928721,
      "grad_norm": 0.4915275275707245,
      "learning_rate": 4.035294316459271e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 8989816,
      "step": 13760
    },
    {
      "epoch": 7.214360587002097,
      "grad_norm": 0.3607673943042755,
      "learning_rate": 4.034391745640215e-05,
      "loss": 0.2935,
      "num_input_tokens_seen": 8992600,
      "step": 13765
    },
    {
      "epoch": 7.216981132075472,
      "grad_norm": 0.423503041267395,
      "learning_rate": 4.033488853843913e-05,
      "loss": 0.472,
      "num_input_tokens_seen": 8997432,
      "step": 13770
    },
    {
      "epoch": 7.219601677148847,
      "grad_norm": 0.6591280102729797,
      "learning_rate": 4.0325856412592363e-05,
      "loss": 0.4966,
      "num_input_tokens_seen": 8999864,
      "step": 13775
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 0.3424873650074005,
      "learning_rate": 4.031682108075129e-05,
      "loss": 0.3818,
      "num_input_tokens_seen": 9003160,
      "step": 13780
    },
    {
      "epoch": 7.2248427672955975,
      "grad_norm": 0.6501588821411133,
      "learning_rate": 4.030778254480599e-05,
      "loss": 0.4495,
      "num_input_tokens_seen": 9006040,
      "step": 13785
    },
    {
      "epoch": 7.227463312368973,
      "grad_norm": 0.44601622223854065,
      "learning_rate": 4.029874080664722e-05,
      "loss": 0.4324,
      "num_input_tokens_seen": 9009624,
      "step": 13790
    },
    {
      "epoch": 7.230083857442348,
      "grad_norm": 0.3719327449798584,
      "learning_rate": 4.0289695868166414e-05,
      "loss": 0.4491,
      "num_input_tokens_seen": 9013912,
      "step": 13795
    },
    {
      "epoch": 7.232704402515723,
      "grad_norm": 0.4044848680496216,
      "learning_rate": 4.028064773125566e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 9017016,
      "step": 13800
    },
    {
      "epoch": 7.235324947589098,
      "grad_norm": 0.7771369814872742,
      "learning_rate": 4.0271596397807734e-05,
      "loss": 0.4052,
      "num_input_tokens_seen": 9019000,
      "step": 13805
    },
    {
      "epoch": 7.237945492662474,
      "grad_norm": 0.5900693535804749,
      "learning_rate": 4.026254186971607e-05,
      "loss": 0.5365,
      "num_input_tokens_seen": 9022520,
      "step": 13810
    },
    {
      "epoch": 7.240566037735849,
      "grad_norm": 0.45220741629600525,
      "learning_rate": 4.025348414887477e-05,
      "loss": 0.4749,
      "num_input_tokens_seen": 9025496,
      "step": 13815
    },
    {
      "epoch": 7.243186582809225,
      "grad_norm": 0.42992478609085083,
      "learning_rate": 4.024442323717861e-05,
      "loss": 0.4502,
      "num_input_tokens_seen": 9029464,
      "step": 13820
    },
    {
      "epoch": 7.2458071278826,
      "grad_norm": 0.49684420228004456,
      "learning_rate": 4.023535913652302e-05,
      "loss": 0.5206,
      "num_input_tokens_seen": 9032088,
      "step": 13825
    },
    {
      "epoch": 7.248427672955975,
      "grad_norm": 0.39813661575317383,
      "learning_rate": 4.022629184880412e-05,
      "loss": 0.4136,
      "num_input_tokens_seen": 9035768,
      "step": 13830
    },
    {
      "epoch": 7.25104821802935,
      "grad_norm": 0.44206610321998596,
      "learning_rate": 4.021722137591867e-05,
      "loss": 0.4385,
      "num_input_tokens_seen": 9038840,
      "step": 13835
    },
    {
      "epoch": 7.253668763102725,
      "grad_norm": 0.4778168201446533,
      "learning_rate": 4.020814771976412e-05,
      "loss": 0.3345,
      "num_input_tokens_seen": 9042328,
      "step": 13840
    },
    {
      "epoch": 7.2562893081761,
      "grad_norm": 0.8565815091133118,
      "learning_rate": 4.019907088223857e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 9045336,
      "step": 13845
    },
    {
      "epoch": 7.258909853249476,
      "grad_norm": 0.4710957705974579,
      "learning_rate": 4.018999086524079e-05,
      "loss": 0.5318,
      "num_input_tokens_seen": 9048312,
      "step": 13850
    },
    {
      "epoch": 7.261530398322851,
      "grad_norm": 0.3956446945667267,
      "learning_rate": 4.018090767067022e-05,
      "loss": 0.5763,
      "num_input_tokens_seen": 9052760,
      "step": 13855
    },
    {
      "epoch": 7.264150943396227,
      "grad_norm": 0.9682471752166748,
      "learning_rate": 4.017182130042696e-05,
      "loss": 0.6801,
      "num_input_tokens_seen": 9055864,
      "step": 13860
    },
    {
      "epoch": 7.266771488469602,
      "grad_norm": 0.7180333733558655,
      "learning_rate": 4.0162731756411764e-05,
      "loss": 0.4625,
      "num_input_tokens_seen": 9058776,
      "step": 13865
    },
    {
      "epoch": 7.269392033542977,
      "grad_norm": 0.5486658811569214,
      "learning_rate": 4.015363904052608e-05,
      "loss": 0.5748,
      "num_input_tokens_seen": 9062360,
      "step": 13870
    },
    {
      "epoch": 7.272012578616352,
      "grad_norm": 0.3029904067516327,
      "learning_rate": 4.014454315467199e-05,
      "loss": 0.5764,
      "num_input_tokens_seen": 9065016,
      "step": 13875
    },
    {
      "epoch": 7.2746331236897275,
      "grad_norm": 0.4044618010520935,
      "learning_rate": 4.013544410075224e-05,
      "loss": 0.4889,
      "num_input_tokens_seen": 9071992,
      "step": 13880
    },
    {
      "epoch": 7.277253668763103,
      "grad_norm": 0.35714301466941833,
      "learning_rate": 4.0126341880670273e-05,
      "loss": 0.4832,
      "num_input_tokens_seen": 9075672,
      "step": 13885
    },
    {
      "epoch": 7.279874213836478,
      "grad_norm": 0.5344204306602478,
      "learning_rate": 4.011723649633014e-05,
      "loss": 0.4192,
      "num_input_tokens_seen": 9078936,
      "step": 13890
    },
    {
      "epoch": 7.282494758909853,
      "grad_norm": 0.6171300411224365,
      "learning_rate": 4.0108127949636606e-05,
      "loss": 0.6314,
      "num_input_tokens_seen": 9082168,
      "step": 13895
    },
    {
      "epoch": 7.285115303983228,
      "grad_norm": 0.718453586101532,
      "learning_rate": 4.009901624249507e-05,
      "loss": 0.6033,
      "num_input_tokens_seen": 9084664,
      "step": 13900
    },
    {
      "epoch": 7.287735849056604,
      "grad_norm": 0.27790772914886475,
      "learning_rate": 4.008990137681159e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 9087864,
      "step": 13905
    },
    {
      "epoch": 7.290356394129979,
      "grad_norm": 0.4862159490585327,
      "learning_rate": 4.00807833544929e-05,
      "loss": 0.5498,
      "num_input_tokens_seen": 9091416,
      "step": 13910
    },
    {
      "epoch": 7.2929769392033545,
      "grad_norm": 0.6659635305404663,
      "learning_rate": 4.0071662177446376e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 9094296,
      "step": 13915
    },
    {
      "epoch": 7.29559748427673,
      "grad_norm": 0.5600412487983704,
      "learning_rate": 4.006253784758008e-05,
      "loss": 0.5231,
      "num_input_tokens_seen": 9097112,
      "step": 13920
    },
    {
      "epoch": 7.298218029350105,
      "grad_norm": 0.39164867997169495,
      "learning_rate": 4.0053410366802704e-05,
      "loss": 0.4408,
      "num_input_tokens_seen": 9100632,
      "step": 13925
    },
    {
      "epoch": 7.30083857442348,
      "grad_norm": 0.4596366882324219,
      "learning_rate": 4.004427973702362e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 9103864,
      "step": 13930
    },
    {
      "epoch": 7.303459119496855,
      "grad_norm": 0.5223968029022217,
      "learning_rate": 4.003514596015285e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 9107928,
      "step": 13935
    },
    {
      "epoch": 7.30607966457023,
      "grad_norm": 0.5004832148551941,
      "learning_rate": 4.002600903810108e-05,
      "loss": 0.3751,
      "num_input_tokens_seen": 9111384,
      "step": 13940
    },
    {
      "epoch": 7.308700209643606,
      "grad_norm": 0.4284128248691559,
      "learning_rate": 4.001686897277963e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 9114488,
      "step": 13945
    },
    {
      "epoch": 7.311320754716981,
      "grad_norm": 0.34110429883003235,
      "learning_rate": 4.000772576610052e-05,
      "loss": 0.4833,
      "num_input_tokens_seen": 9117208,
      "step": 13950
    },
    {
      "epoch": 7.313941299790357,
      "grad_norm": 0.47491762042045593,
      "learning_rate": 3.99985794199764e-05,
      "loss": 0.4665,
      "num_input_tokens_seen": 9121176,
      "step": 13955
    },
    {
      "epoch": 7.316561844863732,
      "grad_norm": 0.4261339008808136,
      "learning_rate": 3.998942993632058e-05,
      "loss": 0.5362,
      "num_input_tokens_seen": 9124216,
      "step": 13960
    },
    {
      "epoch": 7.319182389937107,
      "grad_norm": 0.40718117356300354,
      "learning_rate": 3.998027731704702e-05,
      "loss": 0.4798,
      "num_input_tokens_seen": 9126936,
      "step": 13965
    },
    {
      "epoch": 7.321802935010482,
      "grad_norm": 0.37697434425354004,
      "learning_rate": 3.9971121564070354e-05,
      "loss": 0.5173,
      "num_input_tokens_seen": 9129688,
      "step": 13970
    },
    {
      "epoch": 7.3244234800838575,
      "grad_norm": 0.5088297724723816,
      "learning_rate": 3.996196267930585e-05,
      "loss": 0.4122,
      "num_input_tokens_seen": 9132152,
      "step": 13975
    },
    {
      "epoch": 7.327044025157233,
      "grad_norm": 0.49758467078208923,
      "learning_rate": 3.995280066466945e-05,
      "loss": 0.543,
      "num_input_tokens_seen": 9135352,
      "step": 13980
    },
    {
      "epoch": 7.329664570230608,
      "grad_norm": 0.29255250096321106,
      "learning_rate": 3.9943635522077746e-05,
      "loss": 0.6019,
      "num_input_tokens_seen": 9139064,
      "step": 13985
    },
    {
      "epoch": 7.332285115303983,
      "grad_norm": 0.44330114126205444,
      "learning_rate": 3.993446725344797e-05,
      "loss": 0.4006,
      "num_input_tokens_seen": 9141624,
      "step": 13990
    },
    {
      "epoch": 7.334905660377358,
      "grad_norm": 0.39279794692993164,
      "learning_rate": 3.992529586069803e-05,
      "loss": 0.3801,
      "num_input_tokens_seen": 9144120,
      "step": 13995
    },
    {
      "epoch": 7.337526205450734,
      "grad_norm": 0.6928818821907043,
      "learning_rate": 3.991612134574648e-05,
      "loss": 0.4121,
      "num_input_tokens_seen": 9147384,
      "step": 14000
    },
    {
      "epoch": 7.340146750524109,
      "grad_norm": 0.28713327646255493,
      "learning_rate": 3.99069437105125e-05,
      "loss": 0.4268,
      "num_input_tokens_seen": 9150744,
      "step": 14005
    },
    {
      "epoch": 7.3427672955974845,
      "grad_norm": 0.36307892203330994,
      "learning_rate": 3.9897762956915976e-05,
      "loss": 0.4549,
      "num_input_tokens_seen": 9153560,
      "step": 14010
    },
    {
      "epoch": 7.34538784067086,
      "grad_norm": 0.4018422067165375,
      "learning_rate": 3.9888579086877403e-05,
      "loss": 0.3293,
      "num_input_tokens_seen": 9156696,
      "step": 14015
    },
    {
      "epoch": 7.348008385744235,
      "grad_norm": 0.3878132998943329,
      "learning_rate": 3.987939210231794e-05,
      "loss": 0.3844,
      "num_input_tokens_seen": 9159416,
      "step": 14020
    },
    {
      "epoch": 7.35062893081761,
      "grad_norm": 0.3925398886203766,
      "learning_rate": 3.987020200515941e-05,
      "loss": 0.3515,
      "num_input_tokens_seen": 9162584,
      "step": 14025
    },
    {
      "epoch": 7.353249475890985,
      "grad_norm": 0.4187864065170288,
      "learning_rate": 3.9861008797324255e-05,
      "loss": 0.431,
      "num_input_tokens_seen": 9165848,
      "step": 14030
    },
    {
      "epoch": 7.35587002096436,
      "grad_norm": 0.5183925032615662,
      "learning_rate": 3.9851812480735605e-05,
      "loss": 0.4046,
      "num_input_tokens_seen": 9168600,
      "step": 14035
    },
    {
      "epoch": 7.3584905660377355,
      "grad_norm": 0.33613014221191406,
      "learning_rate": 3.9842613057317235e-05,
      "loss": 0.3799,
      "num_input_tokens_seen": 9173336,
      "step": 14040
    },
    {
      "epoch": 7.361111111111111,
      "grad_norm": 0.47925183176994324,
      "learning_rate": 3.9833410528993534e-05,
      "loss": 0.4682,
      "num_input_tokens_seen": 9176376,
      "step": 14045
    },
    {
      "epoch": 7.363731656184487,
      "grad_norm": 0.441800594329834,
      "learning_rate": 3.9824204897689585e-05,
      "loss": 0.4824,
      "num_input_tokens_seen": 9180088,
      "step": 14050
    },
    {
      "epoch": 7.366352201257862,
      "grad_norm": 1.2896558046340942,
      "learning_rate": 3.98149961653311e-05,
      "loss": 0.4075,
      "num_input_tokens_seen": 9183064,
      "step": 14055
    },
    {
      "epoch": 7.368972746331237,
      "grad_norm": 0.5952346324920654,
      "learning_rate": 3.980578433384442e-05,
      "loss": 0.4372,
      "num_input_tokens_seen": 9186040,
      "step": 14060
    },
    {
      "epoch": 7.371593291404612,
      "grad_norm": 0.4677599370479584,
      "learning_rate": 3.979656940515659e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 9189080,
      "step": 14065
    },
    {
      "epoch": 7.3742138364779874,
      "grad_norm": 0.30539748072624207,
      "learning_rate": 3.9787351381195235e-05,
      "loss": 0.5752,
      "num_input_tokens_seen": 9191992,
      "step": 14070
    },
    {
      "epoch": 7.376834381551363,
      "grad_norm": 0.58623868227005,
      "learning_rate": 3.977813026388868e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 9194520,
      "step": 14075
    },
    {
      "epoch": 7.379454926624738,
      "grad_norm": 0.7398127317428589,
      "learning_rate": 3.976890605516587e-05,
      "loss": 0.4687,
      "num_input_tokens_seen": 9196632,
      "step": 14080
    },
    {
      "epoch": 7.382075471698113,
      "grad_norm": 0.5808908343315125,
      "learning_rate": 3.97596787569564e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 9199832,
      "step": 14085
    },
    {
      "epoch": 7.384696016771488,
      "grad_norm": 0.6258584260940552,
      "learning_rate": 3.975044837119053e-05,
      "loss": 0.4545,
      "num_input_tokens_seen": 9203384,
      "step": 14090
    },
    {
      "epoch": 7.387316561844864,
      "grad_norm": 0.9192889928817749,
      "learning_rate": 3.974121489979914e-05,
      "loss": 0.5002,
      "num_input_tokens_seen": 9206840,
      "step": 14095
    },
    {
      "epoch": 7.389937106918239,
      "grad_norm": 0.6919823288917542,
      "learning_rate": 3.973197834471377e-05,
      "loss": 0.6061,
      "num_input_tokens_seen": 9209432,
      "step": 14100
    },
    {
      "epoch": 7.3925576519916145,
      "grad_norm": 0.3598320484161377,
      "learning_rate": 3.9722738707866595e-05,
      "loss": 0.4352,
      "num_input_tokens_seen": 9212056,
      "step": 14105
    },
    {
      "epoch": 7.39517819706499,
      "grad_norm": 0.3789539039134979,
      "learning_rate": 3.971349599119046e-05,
      "loss": 0.3636,
      "num_input_tokens_seen": 9215224,
      "step": 14110
    },
    {
      "epoch": 7.397798742138365,
      "grad_norm": 0.2722627520561218,
      "learning_rate": 3.970425019661881e-05,
      "loss": 0.3728,
      "num_input_tokens_seen": 9219672,
      "step": 14115
    },
    {
      "epoch": 7.40041928721174,
      "grad_norm": 0.6646470427513123,
      "learning_rate": 3.9695001326085785e-05,
      "loss": 0.4586,
      "num_input_tokens_seen": 9222616,
      "step": 14120
    },
    {
      "epoch": 7.403039832285115,
      "grad_norm": 0.4529602825641632,
      "learning_rate": 3.968574938152613e-05,
      "loss": 0.6906,
      "num_input_tokens_seen": 9227032,
      "step": 14125
    },
    {
      "epoch": 7.40566037735849,
      "grad_norm": 1.1585952043533325,
      "learning_rate": 3.9676494364875247e-05,
      "loss": 0.5149,
      "num_input_tokens_seen": 9229688,
      "step": 14130
    },
    {
      "epoch": 7.4082809224318655,
      "grad_norm": 0.4651867151260376,
      "learning_rate": 3.966723627806918e-05,
      "loss": 0.4984,
      "num_input_tokens_seen": 9233016,
      "step": 14135
    },
    {
      "epoch": 7.410901467505241,
      "grad_norm": 0.4504877030849457,
      "learning_rate": 3.9657975123044624e-05,
      "loss": 0.4497,
      "num_input_tokens_seen": 9236504,
      "step": 14140
    },
    {
      "epoch": 7.413522012578617,
      "grad_norm": 0.45913833379745483,
      "learning_rate": 3.964871090173888e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 9240344,
      "step": 14145
    },
    {
      "epoch": 7.416142557651992,
      "grad_norm": 0.5046811103820801,
      "learning_rate": 3.963944361608995e-05,
      "loss": 0.4441,
      "num_input_tokens_seen": 9242904,
      "step": 14150
    },
    {
      "epoch": 7.418763102725367,
      "grad_norm": 0.7474656701087952,
      "learning_rate": 3.963017326803644e-05,
      "loss": 0.5676,
      "num_input_tokens_seen": 9246072,
      "step": 14155
    },
    {
      "epoch": 7.421383647798742,
      "grad_norm": 0.9000231623649597,
      "learning_rate": 3.962089985951758e-05,
      "loss": 0.4731,
      "num_input_tokens_seen": 9248696,
      "step": 14160
    },
    {
      "epoch": 7.424004192872117,
      "grad_norm": 0.3917349576950073,
      "learning_rate": 3.961162339247327e-05,
      "loss": 0.6269,
      "num_input_tokens_seen": 9251160,
      "step": 14165
    },
    {
      "epoch": 7.426624737945493,
      "grad_norm": 0.4094303548336029,
      "learning_rate": 3.9602343868844055e-05,
      "loss": 0.4983,
      "num_input_tokens_seen": 9254040,
      "step": 14170
    },
    {
      "epoch": 7.429245283018868,
      "grad_norm": 0.2947499454021454,
      "learning_rate": 3.959306129057109e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 9257816,
      "step": 14175
    },
    {
      "epoch": 7.431865828092243,
      "grad_norm": 0.5681163668632507,
      "learning_rate": 3.958377565959619e-05,
      "loss": 0.3897,
      "num_input_tokens_seen": 9260184,
      "step": 14180
    },
    {
      "epoch": 7.434486373165618,
      "grad_norm": 0.5851946473121643,
      "learning_rate": 3.95744869778618e-05,
      "loss": 0.5703,
      "num_input_tokens_seen": 9263032,
      "step": 14185
    },
    {
      "epoch": 7.437106918238994,
      "grad_norm": 0.4406287372112274,
      "learning_rate": 3.956519524731101e-05,
      "loss": 0.5385,
      "num_input_tokens_seen": 9267448,
      "step": 14190
    },
    {
      "epoch": 7.439727463312369,
      "grad_norm": 0.5562689304351807,
      "learning_rate": 3.9555900469887555e-05,
      "loss": 0.4295,
      "num_input_tokens_seen": 9270200,
      "step": 14195
    },
    {
      "epoch": 7.4423480083857445,
      "grad_norm": 0.5893352627754211,
      "learning_rate": 3.954660264753577e-05,
      "loss": 0.4455,
      "num_input_tokens_seen": 9272856,
      "step": 14200
    },
    {
      "epoch": 7.44496855345912,
      "grad_norm": 0.30706116557121277,
      "learning_rate": 3.953730178220067e-05,
      "loss": 0.4267,
      "num_input_tokens_seen": 9276504,
      "step": 14205
    },
    {
      "epoch": 7.447589098532495,
      "grad_norm": 0.47230881452560425,
      "learning_rate": 3.952799787582789e-05,
      "loss": 0.517,
      "num_input_tokens_seen": 9279736,
      "step": 14210
    },
    {
      "epoch": 7.45020964360587,
      "grad_norm": 0.630103349685669,
      "learning_rate": 3.9518690930363694e-05,
      "loss": 0.4069,
      "num_input_tokens_seen": 9282680,
      "step": 14215
    },
    {
      "epoch": 7.452830188679245,
      "grad_norm": 0.7018371224403381,
      "learning_rate": 3.9509380947754994e-05,
      "loss": 0.4499,
      "num_input_tokens_seen": 9285912,
      "step": 14220
    },
    {
      "epoch": 7.45545073375262,
      "grad_norm": 0.5452864170074463,
      "learning_rate": 3.9500067929949336e-05,
      "loss": 0.3612,
      "num_input_tokens_seen": 9289176,
      "step": 14225
    },
    {
      "epoch": 7.4580712788259955,
      "grad_norm": 0.6994613409042358,
      "learning_rate": 3.9490751878894895e-05,
      "loss": 0.4305,
      "num_input_tokens_seen": 9291736,
      "step": 14230
    },
    {
      "epoch": 7.460691823899371,
      "grad_norm": 0.46544262766838074,
      "learning_rate": 3.948143279654047e-05,
      "loss": 0.4414,
      "num_input_tokens_seen": 9296312,
      "step": 14235
    },
    {
      "epoch": 7.463312368972747,
      "grad_norm": 0.7938664555549622,
      "learning_rate": 3.947211068483553e-05,
      "loss": 0.3989,
      "num_input_tokens_seen": 9298904,
      "step": 14240
    },
    {
      "epoch": 7.465932914046122,
      "grad_norm": 0.39895763993263245,
      "learning_rate": 3.946278554573013e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 9301432,
      "step": 14245
    },
    {
      "epoch": 7.468553459119497,
      "grad_norm": 0.41624248027801514,
      "learning_rate": 3.9453457381175e-05,
      "loss": 0.47,
      "num_input_tokens_seen": 9304376,
      "step": 14250
    },
    {
      "epoch": 7.471174004192872,
      "grad_norm": 0.5350850820541382,
      "learning_rate": 3.9444126193121474e-05,
      "loss": 0.4495,
      "num_input_tokens_seen": 9307864,
      "step": 14255
    },
    {
      "epoch": 7.473794549266247,
      "grad_norm": 0.4906122088432312,
      "learning_rate": 3.943479198352153e-05,
      "loss": 0.5077,
      "num_input_tokens_seen": 9311032,
      "step": 14260
    },
    {
      "epoch": 7.476415094339623,
      "grad_norm": 0.3098124861717224,
      "learning_rate": 3.94254547543278e-05,
      "loss": 0.3493,
      "num_input_tokens_seen": 9314264,
      "step": 14265
    },
    {
      "epoch": 7.479035639412998,
      "grad_norm": 0.2731423079967499,
      "learning_rate": 3.9416114507493504e-05,
      "loss": 0.3822,
      "num_input_tokens_seen": 9317272,
      "step": 14270
    },
    {
      "epoch": 7.481656184486373,
      "grad_norm": 0.7579737305641174,
      "learning_rate": 3.940677124497252e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 9320504,
      "step": 14275
    },
    {
      "epoch": 7.484276729559748,
      "grad_norm": 0.4491024911403656,
      "learning_rate": 3.939742496871935e-05,
      "loss": 0.4414,
      "num_input_tokens_seen": 9324792,
      "step": 14280
    },
    {
      "epoch": 7.486897274633124,
      "grad_norm": 0.2847810387611389,
      "learning_rate": 3.9388075680689136e-05,
      "loss": 0.3871,
      "num_input_tokens_seen": 9327544,
      "step": 14285
    },
    {
      "epoch": 7.489517819706499,
      "grad_norm": 0.5009744763374329,
      "learning_rate": 3.9378723382837634e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 9330488,
      "step": 14290
    },
    {
      "epoch": 7.4921383647798745,
      "grad_norm": 0.5841537117958069,
      "learning_rate": 3.936936807712124e-05,
      "loss": 0.5251,
      "num_input_tokens_seen": 9335288,
      "step": 14295
    },
    {
      "epoch": 7.49475890985325,
      "grad_norm": 0.2934113144874573,
      "learning_rate": 3.936000976549698e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 9339384,
      "step": 14300
    },
    {
      "epoch": 7.497379454926625,
      "grad_norm": 0.38842934370040894,
      "learning_rate": 3.935064844992251e-05,
      "loss": 0.5714,
      "num_input_tokens_seen": 9342872,
      "step": 14305
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.3816433250904083,
      "learning_rate": 3.93412841323561e-05,
      "loss": 0.4093,
      "num_input_tokens_seen": 9345976,
      "step": 14310
    },
    {
      "epoch": 7.502620545073375,
      "grad_norm": 0.5044752359390259,
      "learning_rate": 3.933191681475666e-05,
      "loss": 0.4072,
      "num_input_tokens_seen": 9349208,
      "step": 14315
    },
    {
      "epoch": 7.50524109014675,
      "grad_norm": 0.5300202965736389,
      "learning_rate": 3.932254649908373e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 9352184,
      "step": 14320
    },
    {
      "epoch": 7.5078616352201255,
      "grad_norm": 0.48306789994239807,
      "learning_rate": 3.931317318729748e-05,
      "loss": 0.4929,
      "num_input_tokens_seen": 9356216,
      "step": 14325
    },
    {
      "epoch": 7.510482180293501,
      "grad_norm": 0.6976734399795532,
      "learning_rate": 3.930379688135869e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 9359544,
      "step": 14330
    },
    {
      "epoch": 7.513102725366876,
      "grad_norm": 0.32310956716537476,
      "learning_rate": 3.929441758322878e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 9362200,
      "step": 14335
    },
    {
      "epoch": 7.515723270440252,
      "grad_norm": 0.36432597041130066,
      "learning_rate": 3.928503529486979e-05,
      "loss": 0.4065,
      "num_input_tokens_seen": 9366520,
      "step": 14340
    },
    {
      "epoch": 7.518343815513627,
      "grad_norm": 0.31620484590530396,
      "learning_rate": 3.927565001824438e-05,
      "loss": 0.3889,
      "num_input_tokens_seen": 9372440,
      "step": 14345
    },
    {
      "epoch": 7.520964360587002,
      "grad_norm": 0.5613359212875366,
      "learning_rate": 3.926626175531586e-05,
      "loss": 0.3564,
      "num_input_tokens_seen": 9375064,
      "step": 14350
    },
    {
      "epoch": 7.523584905660377,
      "grad_norm": 0.5573669075965881,
      "learning_rate": 3.925687050804814e-05,
      "loss": 0.3713,
      "num_input_tokens_seen": 9377720,
      "step": 14355
    },
    {
      "epoch": 7.526205450733753,
      "grad_norm": 0.3520212173461914,
      "learning_rate": 3.924747627840575e-05,
      "loss": 0.4641,
      "num_input_tokens_seen": 9380984,
      "step": 14360
    },
    {
      "epoch": 7.528825995807128,
      "grad_norm": 0.5828385949134827,
      "learning_rate": 3.923807906835388e-05,
      "loss": 0.3885,
      "num_input_tokens_seen": 9383992,
      "step": 14365
    },
    {
      "epoch": 7.531446540880503,
      "grad_norm": 1.083922028541565,
      "learning_rate": 3.92286788798583e-05,
      "loss": 0.4637,
      "num_input_tokens_seen": 9386360,
      "step": 14370
    },
    {
      "epoch": 7.534067085953878,
      "grad_norm": 0.6524122953414917,
      "learning_rate": 3.921927571488543e-05,
      "loss": 0.5206,
      "num_input_tokens_seen": 9389144,
      "step": 14375
    },
    {
      "epoch": 7.536687631027254,
      "grad_norm": 0.6027932167053223,
      "learning_rate": 3.9209869575402296e-05,
      "loss": 0.3411,
      "num_input_tokens_seen": 9392024,
      "step": 14380
    },
    {
      "epoch": 7.539308176100629,
      "grad_norm": 0.7485069036483765,
      "learning_rate": 3.9200460463376566e-05,
      "loss": 0.4366,
      "num_input_tokens_seen": 9394872,
      "step": 14385
    },
    {
      "epoch": 7.5419287211740045,
      "grad_norm": 0.4488903284072876,
      "learning_rate": 3.9191048380776515e-05,
      "loss": 0.6201,
      "num_input_tokens_seen": 9397304,
      "step": 14390
    },
    {
      "epoch": 7.54454926624738,
      "grad_norm": 0.5313181281089783,
      "learning_rate": 3.918163332957104e-05,
      "loss": 0.5065,
      "num_input_tokens_seen": 9400664,
      "step": 14395
    },
    {
      "epoch": 7.547169811320755,
      "grad_norm": 0.2950040102005005,
      "learning_rate": 3.9172215311729665e-05,
      "loss": 0.5227,
      "num_input_tokens_seen": 9403864,
      "step": 14400
    },
    {
      "epoch": 7.54979035639413,
      "grad_norm": 0.7564597129821777,
      "learning_rate": 3.916279432922254e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 9407544,
      "step": 14405
    },
    {
      "epoch": 7.552410901467505,
      "grad_norm": 0.20037832856178284,
      "learning_rate": 3.915337038402041e-05,
      "loss": 0.3594,
      "num_input_tokens_seen": 9410968,
      "step": 14410
    },
    {
      "epoch": 7.55503144654088,
      "grad_norm": 0.6877024173736572,
      "learning_rate": 3.914394347809467e-05,
      "loss": 0.5022,
      "num_input_tokens_seen": 9413400,
      "step": 14415
    },
    {
      "epoch": 7.5576519916142555,
      "grad_norm": 0.6807935237884521,
      "learning_rate": 3.9134513613417315e-05,
      "loss": 0.368,
      "num_input_tokens_seen": 9415832,
      "step": 14420
    },
    {
      "epoch": 7.560272536687631,
      "grad_norm": 0.5001965165138245,
      "learning_rate": 3.912508079196097e-05,
      "loss": 0.5521,
      "num_input_tokens_seen": 9419064,
      "step": 14425
    },
    {
      "epoch": 7.562893081761006,
      "grad_norm": 0.43304330110549927,
      "learning_rate": 3.9115645015698866e-05,
      "loss": 0.505,
      "num_input_tokens_seen": 9421752,
      "step": 14430
    },
    {
      "epoch": 7.565513626834382,
      "grad_norm": 0.5305205583572388,
      "learning_rate": 3.9106206286604865e-05,
      "loss": 0.5036,
      "num_input_tokens_seen": 9424984,
      "step": 14435
    },
    {
      "epoch": 7.568134171907757,
      "grad_norm": 0.444296270608902,
      "learning_rate": 3.909676460665345e-05,
      "loss": 0.4,
      "num_input_tokens_seen": 9428216,
      "step": 14440
    },
    {
      "epoch": 7.570754716981132,
      "grad_norm": 0.5379127860069275,
      "learning_rate": 3.908731997781969e-05,
      "loss": 0.4657,
      "num_input_tokens_seen": 9431576,
      "step": 14445
    },
    {
      "epoch": 7.573375262054507,
      "grad_norm": 0.685133159160614,
      "learning_rate": 3.907787240207932e-05,
      "loss": 0.5895,
      "num_input_tokens_seen": 9434680,
      "step": 14450
    },
    {
      "epoch": 7.575995807127883,
      "grad_norm": 0.7471866011619568,
      "learning_rate": 3.9068421881408645e-05,
      "loss": 0.2856,
      "num_input_tokens_seen": 9437432,
      "step": 14455
    },
    {
      "epoch": 7.578616352201258,
      "grad_norm": 0.38528311252593994,
      "learning_rate": 3.9058968417784614e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 9441208,
      "step": 14460
    },
    {
      "epoch": 7.581236897274633,
      "grad_norm": 0.5833650231361389,
      "learning_rate": 3.9049512013184785e-05,
      "loss": 0.4608,
      "num_input_tokens_seen": 9444056,
      "step": 14465
    },
    {
      "epoch": 7.583857442348008,
      "grad_norm": 1.8695751428604126,
      "learning_rate": 3.9040052669587325e-05,
      "loss": 0.5652,
      "num_input_tokens_seen": 9446968,
      "step": 14470
    },
    {
      "epoch": 7.586477987421384,
      "grad_norm": 0.3263431191444397,
      "learning_rate": 3.9030590388971026e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 9450712,
      "step": 14475
    },
    {
      "epoch": 7.589098532494759,
      "grad_norm": 0.3149089515209198,
      "learning_rate": 3.902112517331529e-05,
      "loss": 0.4175,
      "num_input_tokens_seen": 9453240,
      "step": 14480
    },
    {
      "epoch": 7.5917190775681345,
      "grad_norm": 0.4079769253730774,
      "learning_rate": 3.901165702460012e-05,
      "loss": 0.385,
      "num_input_tokens_seen": 9456728,
      "step": 14485
    },
    {
      "epoch": 7.59433962264151,
      "grad_norm": 0.4500882029533386,
      "learning_rate": 3.900218594480616e-05,
      "loss": 0.454,
      "num_input_tokens_seen": 9459800,
      "step": 14490
    },
    {
      "epoch": 7.596960167714885,
      "grad_norm": 0.38484469056129456,
      "learning_rate": 3.899271193591465e-05,
      "loss": 0.4317,
      "num_input_tokens_seen": 9463384,
      "step": 14495
    },
    {
      "epoch": 7.59958071278826,
      "grad_norm": 0.4599692225456238,
      "learning_rate": 3.8983234999907435e-05,
      "loss": 0.4504,
      "num_input_tokens_seen": 9466520,
      "step": 14500
    },
    {
      "epoch": 7.602201257861635,
      "grad_norm": 0.41257697343826294,
      "learning_rate": 3.897375513876699e-05,
      "loss": 0.3683,
      "num_input_tokens_seen": 9469944,
      "step": 14505
    },
    {
      "epoch": 7.60482180293501,
      "grad_norm": 2.8311197757720947,
      "learning_rate": 3.8964272354476396e-05,
      "loss": 0.4949,
      "num_input_tokens_seen": 9472280,
      "step": 14510
    },
    {
      "epoch": 7.6074423480083855,
      "grad_norm": 0.6355085372924805,
      "learning_rate": 3.895478664901933e-05,
      "loss": 0.46,
      "num_input_tokens_seen": 9474808,
      "step": 14515
    },
    {
      "epoch": 7.610062893081761,
      "grad_norm": 0.29393208026885986,
      "learning_rate": 3.89452980243801e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 9479064,
      "step": 14520
    },
    {
      "epoch": 7.612683438155136,
      "grad_norm": 0.38884520530700684,
      "learning_rate": 3.893580648254362e-05,
      "loss": 0.4821,
      "num_input_tokens_seen": 9483128,
      "step": 14525
    },
    {
      "epoch": 7.615303983228512,
      "grad_norm": 0.6077716946601868,
      "learning_rate": 3.8926312025495414e-05,
      "loss": 0.441,
      "num_input_tokens_seen": 9486136,
      "step": 14530
    },
    {
      "epoch": 7.617924528301887,
      "grad_norm": 1.0759391784667969,
      "learning_rate": 3.891681465522161e-05,
      "loss": 0.7952,
      "num_input_tokens_seen": 9488568,
      "step": 14535
    },
    {
      "epoch": 7.620545073375262,
      "grad_norm": 0.6043104529380798,
      "learning_rate": 3.890731437370896e-05,
      "loss": 0.46,
      "num_input_tokens_seen": 9492440,
      "step": 14540
    },
    {
      "epoch": 7.623165618448637,
      "grad_norm": 0.4187376797199249,
      "learning_rate": 3.88978111829448e-05,
      "loss": 0.4098,
      "num_input_tokens_seen": 9495192,
      "step": 14545
    },
    {
      "epoch": 7.6257861635220126,
      "grad_norm": 1.0144031047821045,
      "learning_rate": 3.88883050849171e-05,
      "loss": 0.4897,
      "num_input_tokens_seen": 9497688,
      "step": 14550
    },
    {
      "epoch": 7.628406708595388,
      "grad_norm": 0.3438616394996643,
      "learning_rate": 3.8878796081614417e-05,
      "loss": 0.4527,
      "num_input_tokens_seen": 9500984,
      "step": 14555
    },
    {
      "epoch": 7.631027253668763,
      "grad_norm": 0.3553379476070404,
      "learning_rate": 3.8869284175025935e-05,
      "loss": 0.4341,
      "num_input_tokens_seen": 9504152,
      "step": 14560
    },
    {
      "epoch": 7.633647798742138,
      "grad_norm": 0.3618480861186981,
      "learning_rate": 3.885976936714143e-05,
      "loss": 0.4554,
      "num_input_tokens_seen": 9506840,
      "step": 14565
    },
    {
      "epoch": 7.636268343815514,
      "grad_norm": 0.2846296429634094,
      "learning_rate": 3.885025165995129e-05,
      "loss": 0.4883,
      "num_input_tokens_seen": 9509880,
      "step": 14570
    },
    {
      "epoch": 7.638888888888889,
      "grad_norm": 0.5899193286895752,
      "learning_rate": 3.884073105544652e-05,
      "loss": 0.429,
      "num_input_tokens_seen": 9513464,
      "step": 14575
    },
    {
      "epoch": 7.6415094339622645,
      "grad_norm": 0.8576472997665405,
      "learning_rate": 3.8831207555618723e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 9516216,
      "step": 14580
    },
    {
      "epoch": 7.64412997903564,
      "grad_norm": 0.40733620524406433,
      "learning_rate": 3.882168116246009e-05,
      "loss": 0.3699,
      "num_input_tokens_seen": 9519032,
      "step": 14585
    },
    {
      "epoch": 7.646750524109015,
      "grad_norm": 0.2680160403251648,
      "learning_rate": 3.881215187796344e-05,
      "loss": 0.3729,
      "num_input_tokens_seen": 9523640,
      "step": 14590
    },
    {
      "epoch": 7.64937106918239,
      "grad_norm": 0.4373973608016968,
      "learning_rate": 3.8802619704122204e-05,
      "loss": 0.4177,
      "num_input_tokens_seen": 9526616,
      "step": 14595
    },
    {
      "epoch": 7.651991614255765,
      "grad_norm": 0.37761202454566956,
      "learning_rate": 3.8793084642930387e-05,
      "loss": 0.4744,
      "num_input_tokens_seen": 9529656,
      "step": 14600
    },
    {
      "epoch": 7.65461215932914,
      "grad_norm": 0.2751489579677582,
      "learning_rate": 3.878354669638263e-05,
      "loss": 0.4126,
      "num_input_tokens_seen": 9532984,
      "step": 14605
    },
    {
      "epoch": 7.6572327044025155,
      "grad_norm": 0.8087230324745178,
      "learning_rate": 3.8774005866474155e-05,
      "loss": 0.4231,
      "num_input_tokens_seen": 9535928,
      "step": 14610
    },
    {
      "epoch": 7.659853249475891,
      "grad_norm": 0.450142502784729,
      "learning_rate": 3.876446215520079e-05,
      "loss": 0.4865,
      "num_input_tokens_seen": 9539256,
      "step": 14615
    },
    {
      "epoch": 7.662473794549266,
      "grad_norm": 0.31245627999305725,
      "learning_rate": 3.875491556455898e-05,
      "loss": 0.4475,
      "num_input_tokens_seen": 9542296,
      "step": 14620
    },
    {
      "epoch": 7.665094339622642,
      "grad_norm": 0.4829571843147278,
      "learning_rate": 3.8745366096545754e-05,
      "loss": 0.3673,
      "num_input_tokens_seen": 9545368,
      "step": 14625
    },
    {
      "epoch": 7.667714884696017,
      "grad_norm": 0.2669237554073334,
      "learning_rate": 3.873581375315876e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 9548472,
      "step": 14630
    },
    {
      "epoch": 7.670335429769392,
      "grad_norm": 0.24869000911712646,
      "learning_rate": 3.872625853639623e-05,
      "loss": 0.3893,
      "num_input_tokens_seen": 9552248,
      "step": 14635
    },
    {
      "epoch": 7.672955974842767,
      "grad_norm": 0.4669564664363861,
      "learning_rate": 3.871670044825701e-05,
      "loss": 0.6936,
      "num_input_tokens_seen": 9555320,
      "step": 14640
    },
    {
      "epoch": 7.6755765199161425,
      "grad_norm": 0.5555433034896851,
      "learning_rate": 3.870713949074056e-05,
      "loss": 0.4613,
      "num_input_tokens_seen": 9558904,
      "step": 14645
    },
    {
      "epoch": 7.678197064989518,
      "grad_norm": 0.6992555260658264,
      "learning_rate": 3.86975756658469e-05,
      "loss": 0.4922,
      "num_input_tokens_seen": 9561624,
      "step": 14650
    },
    {
      "epoch": 7.680817610062893,
      "grad_norm": 0.6403408646583557,
      "learning_rate": 3.868800897557667e-05,
      "loss": 0.5856,
      "num_input_tokens_seen": 9564568,
      "step": 14655
    },
    {
      "epoch": 7.683438155136268,
      "grad_norm": 0.498293399810791,
      "learning_rate": 3.867843942193113e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 9567704,
      "step": 14660
    },
    {
      "epoch": 7.686058700209644,
      "grad_norm": 0.435928076505661,
      "learning_rate": 3.866886700691212e-05,
      "loss": 0.45,
      "num_input_tokens_seen": 9571512,
      "step": 14665
    },
    {
      "epoch": 7.688679245283019,
      "grad_norm": 0.42112693190574646,
      "learning_rate": 3.8659291732522075e-05,
      "loss": 0.5464,
      "num_input_tokens_seen": 9575448,
      "step": 14670
    },
    {
      "epoch": 7.691299790356394,
      "grad_norm": 0.3095299005508423,
      "learning_rate": 3.8649713600764026e-05,
      "loss": 0.2747,
      "num_input_tokens_seen": 9583704,
      "step": 14675
    },
    {
      "epoch": 7.69392033542977,
      "grad_norm": 0.4417037069797516,
      "learning_rate": 3.864013261364162e-05,
      "loss": 0.4312,
      "num_input_tokens_seen": 9586392,
      "step": 14680
    },
    {
      "epoch": 7.696540880503145,
      "grad_norm": 0.5111088752746582,
      "learning_rate": 3.8630548773159084e-05,
      "loss": 0.588,
      "num_input_tokens_seen": 9589336,
      "step": 14685
    },
    {
      "epoch": 7.69916142557652,
      "grad_norm": 0.2864827811717987,
      "learning_rate": 3.862096208132125e-05,
      "loss": 0.4146,
      "num_input_tokens_seen": 9593016,
      "step": 14690
    },
    {
      "epoch": 7.701781970649895,
      "grad_norm": 0.6021435856819153,
      "learning_rate": 3.861137254013355e-05,
      "loss": 0.4243,
      "num_input_tokens_seen": 9595512,
      "step": 14695
    },
    {
      "epoch": 7.70440251572327,
      "grad_norm": 0.5725905299186707,
      "learning_rate": 3.8601780151601996e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 9598904,
      "step": 14700
    },
    {
      "epoch": 7.7070230607966455,
      "grad_norm": 0.35435751080513,
      "learning_rate": 3.859218491773321e-05,
      "loss": 0.2936,
      "num_input_tokens_seen": 9602712,
      "step": 14705
    },
    {
      "epoch": 7.709643605870021,
      "grad_norm": 0.3393063247203827,
      "learning_rate": 3.858258684053441e-05,
      "loss": 0.4625,
      "num_input_tokens_seen": 9605560,
      "step": 14710
    },
    {
      "epoch": 7.712264150943396,
      "grad_norm": 0.38712450861930847,
      "learning_rate": 3.8572985922013405e-05,
      "loss": 0.583,
      "num_input_tokens_seen": 9609048,
      "step": 14715
    },
    {
      "epoch": 7.714884696016772,
      "grad_norm": 0.5934461355209351,
      "learning_rate": 3.856338216417859e-05,
      "loss": 0.4373,
      "num_input_tokens_seen": 9612824,
      "step": 14720
    },
    {
      "epoch": 7.717505241090147,
      "grad_norm": 0.2391500025987625,
      "learning_rate": 3.855377556903897e-05,
      "loss": 0.4024,
      "num_input_tokens_seen": 9615704,
      "step": 14725
    },
    {
      "epoch": 7.720125786163522,
      "grad_norm": 0.5849622488021851,
      "learning_rate": 3.8544166138604124e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 9618840,
      "step": 14730
    },
    {
      "epoch": 7.722746331236897,
      "grad_norm": 0.3374795615673065,
      "learning_rate": 3.8534553874884244e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 9622552,
      "step": 14735
    },
    {
      "epoch": 7.7253668763102725,
      "grad_norm": 0.268817275762558,
      "learning_rate": 3.852493877989011e-05,
      "loss": 0.5914,
      "num_input_tokens_seen": 9625720,
      "step": 14740
    },
    {
      "epoch": 7.727987421383648,
      "grad_norm": 0.43009138107299805,
      "learning_rate": 3.851532085563308e-05,
      "loss": 0.3923,
      "num_input_tokens_seen": 9628664,
      "step": 14745
    },
    {
      "epoch": 7.730607966457023,
      "grad_norm": 0.3457311689853668,
      "learning_rate": 3.850570010412511e-05,
      "loss": 0.4077,
      "num_input_tokens_seen": 9632088,
      "step": 14750
    },
    {
      "epoch": 7.733228511530398,
      "grad_norm": 0.448529452085495,
      "learning_rate": 3.8496076527378766e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 9635160,
      "step": 14755
    },
    {
      "epoch": 7.735849056603773,
      "grad_norm": 0.37280696630477905,
      "learning_rate": 3.848645012740718e-05,
      "loss": 0.4734,
      "num_input_tokens_seen": 9638776,
      "step": 14760
    },
    {
      "epoch": 7.738469601677149,
      "grad_norm": 0.8921027779579163,
      "learning_rate": 3.8476820906224096e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 9641624,
      "step": 14765
    },
    {
      "epoch": 7.741090146750524,
      "grad_norm": 0.314965158700943,
      "learning_rate": 3.846718886584382e-05,
      "loss": 0.5271,
      "num_input_tokens_seen": 9649176,
      "step": 14770
    },
    {
      "epoch": 7.7437106918239,
      "grad_norm": 0.4789530634880066,
      "learning_rate": 3.845755400828128e-05,
      "loss": 0.6536,
      "num_input_tokens_seen": 9651768,
      "step": 14775
    },
    {
      "epoch": 7.746331236897275,
      "grad_norm": 0.3699617385864258,
      "learning_rate": 3.844791633555198e-05,
      "loss": 0.2928,
      "num_input_tokens_seen": 9654552,
      "step": 14780
    },
    {
      "epoch": 7.74895178197065,
      "grad_norm": 0.3149905800819397,
      "learning_rate": 3.8438275849671996e-05,
      "loss": 0.4426,
      "num_input_tokens_seen": 9657112,
      "step": 14785
    },
    {
      "epoch": 7.751572327044025,
      "grad_norm": 0.3183784782886505,
      "learning_rate": 3.842863255265802e-05,
      "loss": 0.3823,
      "num_input_tokens_seen": 9661176,
      "step": 14790
    },
    {
      "epoch": 7.7541928721174,
      "grad_norm": 0.5287168025970459,
      "learning_rate": 3.841898644652732e-05,
      "loss": 0.4375,
      "num_input_tokens_seen": 9663480,
      "step": 14795
    },
    {
      "epoch": 7.756813417190775,
      "grad_norm": 0.3024071455001831,
      "learning_rate": 3.840933753329774e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 9666808,
      "step": 14800
    },
    {
      "epoch": 7.759433962264151,
      "grad_norm": 0.7099863886833191,
      "learning_rate": 3.839968581498773e-05,
      "loss": 0.5949,
      "num_input_tokens_seen": 9669336,
      "step": 14805
    },
    {
      "epoch": 7.762054507337526,
      "grad_norm": 0.4612322449684143,
      "learning_rate": 3.839003129361632e-05,
      "loss": 0.5525,
      "num_input_tokens_seen": 9671928,
      "step": 14810
    },
    {
      "epoch": 7.764675052410902,
      "grad_norm": 0.34149906039237976,
      "learning_rate": 3.8380373971203134e-05,
      "loss": 0.5269,
      "num_input_tokens_seen": 9676120,
      "step": 14815
    },
    {
      "epoch": 7.767295597484277,
      "grad_norm": 0.5692480206489563,
      "learning_rate": 3.8370713849768346e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 9679032,
      "step": 14820
    },
    {
      "epoch": 7.769916142557652,
      "grad_norm": 0.47175511717796326,
      "learning_rate": 3.8361050931332774e-05,
      "loss": 0.4362,
      "num_input_tokens_seen": 9682744,
      "step": 14825
    },
    {
      "epoch": 7.772536687631027,
      "grad_norm": 0.4825233221054077,
      "learning_rate": 3.835138521791778e-05,
      "loss": 0.6457,
      "num_input_tokens_seen": 9686040,
      "step": 14830
    },
    {
      "epoch": 7.7751572327044025,
      "grad_norm": 0.5255576372146606,
      "learning_rate": 3.8341716711545314e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 9690168,
      "step": 14835
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 0.6058690547943115,
      "learning_rate": 3.833204541423793e-05,
      "loss": 0.4868,
      "num_input_tokens_seen": 9693496,
      "step": 14840
    },
    {
      "epoch": 7.780398322851153,
      "grad_norm": 0.5541348457336426,
      "learning_rate": 3.832237132801874e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 9696760,
      "step": 14845
    },
    {
      "epoch": 7.783018867924528,
      "grad_norm": 0.38641926646232605,
      "learning_rate": 3.831269445491146e-05,
      "loss": 0.5328,
      "num_input_tokens_seen": 9699672,
      "step": 14850
    },
    {
      "epoch": 7.785639412997903,
      "grad_norm": 0.3275163471698761,
      "learning_rate": 3.8303014796940375e-05,
      "loss": 0.4014,
      "num_input_tokens_seen": 9703384,
      "step": 14855
    },
    {
      "epoch": 7.788259958071279,
      "grad_norm": 0.2498980313539505,
      "learning_rate": 3.829333235613036e-05,
      "loss": 0.5011,
      "num_input_tokens_seen": 9706328,
      "step": 14860
    },
    {
      "epoch": 7.790880503144654,
      "grad_norm": 0.5078223943710327,
      "learning_rate": 3.828364713450688e-05,
      "loss": 0.5283,
      "num_input_tokens_seen": 9709432,
      "step": 14865
    },
    {
      "epoch": 7.79350104821803,
      "grad_norm": 0.7705269455909729,
      "learning_rate": 3.827395913409597e-05,
      "loss": 0.5723,
      "num_input_tokens_seen": 9712696,
      "step": 14870
    },
    {
      "epoch": 7.796121593291405,
      "grad_norm": 0.6191033720970154,
      "learning_rate": 3.826426835692424e-05,
      "loss": 0.5237,
      "num_input_tokens_seen": 9715704,
      "step": 14875
    },
    {
      "epoch": 7.79874213836478,
      "grad_norm": 0.4672381281852722,
      "learning_rate": 3.825457480501891e-05,
      "loss": 0.6793,
      "num_input_tokens_seen": 9720344,
      "step": 14880
    },
    {
      "epoch": 7.801362683438155,
      "grad_norm": 0.28226780891418457,
      "learning_rate": 3.824487848040773e-05,
      "loss": 0.4151,
      "num_input_tokens_seen": 9723544,
      "step": 14885
    },
    {
      "epoch": 7.80398322851153,
      "grad_norm": 0.5196695327758789,
      "learning_rate": 3.823517938511908e-05,
      "loss": 0.6963,
      "num_input_tokens_seen": 9726424,
      "step": 14890
    },
    {
      "epoch": 7.806603773584905,
      "grad_norm": 0.5600076913833618,
      "learning_rate": 3.8225477521181906e-05,
      "loss": 0.4377,
      "num_input_tokens_seen": 9729144,
      "step": 14895
    },
    {
      "epoch": 7.809224318658281,
      "grad_norm": 0.291669636964798,
      "learning_rate": 3.821577289062571e-05,
      "loss": 0.4389,
      "num_input_tokens_seen": 9732504,
      "step": 14900
    },
    {
      "epoch": 7.811844863731656,
      "grad_norm": 0.42647331953048706,
      "learning_rate": 3.82060654954806e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 9737048,
      "step": 14905
    },
    {
      "epoch": 7.814465408805032,
      "grad_norm": 0.6588507890701294,
      "learning_rate": 3.819635533777726e-05,
      "loss": 0.4959,
      "num_input_tokens_seen": 9739960,
      "step": 14910
    },
    {
      "epoch": 7.817085953878407,
      "grad_norm": 0.44039756059646606,
      "learning_rate": 3.8186642419546924e-05,
      "loss": 0.3753,
      "num_input_tokens_seen": 9743288,
      "step": 14915
    },
    {
      "epoch": 7.819706498951782,
      "grad_norm": 0.564334511756897,
      "learning_rate": 3.817692674282145e-05,
      "loss": 0.6802,
      "num_input_tokens_seen": 9746072,
      "step": 14920
    },
    {
      "epoch": 7.822327044025157,
      "grad_norm": 0.39345818758010864,
      "learning_rate": 3.816720830963322e-05,
      "loss": 0.5187,
      "num_input_tokens_seen": 9748920,
      "step": 14925
    },
    {
      "epoch": 7.8249475890985325,
      "grad_norm": 0.31329378485679626,
      "learning_rate": 3.815748712201523e-05,
      "loss": 0.3367,
      "num_input_tokens_seen": 9751864,
      "step": 14930
    },
    {
      "epoch": 7.827568134171908,
      "grad_norm": 0.3397785723209381,
      "learning_rate": 3.814776318200105e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 9755128,
      "step": 14935
    },
    {
      "epoch": 7.830188679245283,
      "grad_norm": 0.9672359228134155,
      "learning_rate": 3.8138036491624805e-05,
      "loss": 0.3655,
      "num_input_tokens_seen": 9757976,
      "step": 14940
    },
    {
      "epoch": 7.832809224318658,
      "grad_norm": 0.26414304971694946,
      "learning_rate": 3.812830705292122e-05,
      "loss": 0.3865,
      "num_input_tokens_seen": 9761464,
      "step": 14945
    },
    {
      "epoch": 7.835429769392033,
      "grad_norm": 0.6283736824989319,
      "learning_rate": 3.8118574867925564e-05,
      "loss": 0.4584,
      "num_input_tokens_seen": 9768408,
      "step": 14950
    },
    {
      "epoch": 7.838050314465409,
      "grad_norm": 0.3253106474876404,
      "learning_rate": 3.810883993867371e-05,
      "loss": 0.5134,
      "num_input_tokens_seen": 9771832,
      "step": 14955
    },
    {
      "epoch": 7.840670859538784,
      "grad_norm": 0.39980649948120117,
      "learning_rate": 3.80991022672021e-05,
      "loss": 0.5568,
      "num_input_tokens_seen": 9775736,
      "step": 14960
    },
    {
      "epoch": 7.84329140461216,
      "grad_norm": 0.40962785482406616,
      "learning_rate": 3.8089361855547744e-05,
      "loss": 0.5158,
      "num_input_tokens_seen": 9779096,
      "step": 14965
    },
    {
      "epoch": 7.845911949685535,
      "grad_norm": 0.7343834638595581,
      "learning_rate": 3.8079618705748214e-05,
      "loss": 0.7222,
      "num_input_tokens_seen": 9782264,
      "step": 14970
    },
    {
      "epoch": 7.84853249475891,
      "grad_norm": 0.3702188730239868,
      "learning_rate": 3.806987281984167e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 9785368,
      "step": 14975
    },
    {
      "epoch": 7.851153039832285,
      "grad_norm": 0.6776047945022583,
      "learning_rate": 3.806012419986684e-05,
      "loss": 0.3832,
      "num_input_tokens_seen": 9789304,
      "step": 14980
    },
    {
      "epoch": 7.85377358490566,
      "grad_norm": 0.561959981918335,
      "learning_rate": 3.805037284786303e-05,
      "loss": 0.5876,
      "num_input_tokens_seen": 9792184,
      "step": 14985
    },
    {
      "epoch": 7.856394129979035,
      "grad_norm": 0.45757031440734863,
      "learning_rate": 3.804061876587009e-05,
      "loss": 0.5493,
      "num_input_tokens_seen": 9795928,
      "step": 14990
    },
    {
      "epoch": 7.859014675052411,
      "grad_norm": 0.45549845695495605,
      "learning_rate": 3.80308619559285e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 9798776,
      "step": 14995
    },
    {
      "epoch": 7.861635220125786,
      "grad_norm": 0.38432422280311584,
      "learning_rate": 3.802110242007925e-05,
      "loss": 0.393,
      "num_input_tokens_seen": 9802008,
      "step": 15000
    },
    {
      "epoch": 7.864255765199162,
      "grad_norm": 0.9877076745033264,
      "learning_rate": 3.801134016036392e-05,
      "loss": 0.6091,
      "num_input_tokens_seen": 9805912,
      "step": 15005
    },
    {
      "epoch": 7.866876310272537,
      "grad_norm": 0.4291321039199829,
      "learning_rate": 3.8001575178824665e-05,
      "loss": 0.4598,
      "num_input_tokens_seen": 9808664,
      "step": 15010
    },
    {
      "epoch": 7.869496855345912,
      "grad_norm": 0.27809783816337585,
      "learning_rate": 3.799180747750422e-05,
      "loss": 0.3993,
      "num_input_tokens_seen": 9811256,
      "step": 15015
    },
    {
      "epoch": 7.872117400419287,
      "grad_norm": 0.34969428181648254,
      "learning_rate": 3.798203705844586e-05,
      "loss": 0.5137,
      "num_input_tokens_seen": 9816440,
      "step": 15020
    },
    {
      "epoch": 7.8747379454926625,
      "grad_norm": 0.39604994654655457,
      "learning_rate": 3.7972263923693476e-05,
      "loss": 0.4407,
      "num_input_tokens_seen": 9818776,
      "step": 15025
    },
    {
      "epoch": 7.877358490566038,
      "grad_norm": 0.44848009943962097,
      "learning_rate": 3.796248807529145e-05,
      "loss": 0.4791,
      "num_input_tokens_seen": 9822520,
      "step": 15030
    },
    {
      "epoch": 7.879979035639413,
      "grad_norm": 0.33735519647598267,
      "learning_rate": 3.795270951528482e-05,
      "loss": 0.3626,
      "num_input_tokens_seen": 9825496,
      "step": 15035
    },
    {
      "epoch": 7.882599580712788,
      "grad_norm": 0.3546415865421295,
      "learning_rate": 3.7942928245719124e-05,
      "loss": 0.6153,
      "num_input_tokens_seen": 9829080,
      "step": 15040
    },
    {
      "epoch": 7.885220125786163,
      "grad_norm": 0.6651219725608826,
      "learning_rate": 3.7933144268640495e-05,
      "loss": 0.4184,
      "num_input_tokens_seen": 9832056,
      "step": 15045
    },
    {
      "epoch": 7.887840670859539,
      "grad_norm": 0.43730056285858154,
      "learning_rate": 3.7923357586095645e-05,
      "loss": 0.4805,
      "num_input_tokens_seen": 9834680,
      "step": 15050
    },
    {
      "epoch": 7.890461215932914,
      "grad_norm": 0.43954697251319885,
      "learning_rate": 3.7913568200131823e-05,
      "loss": 0.5477,
      "num_input_tokens_seen": 9838008,
      "step": 15055
    },
    {
      "epoch": 7.8930817610062896,
      "grad_norm": 0.4730432629585266,
      "learning_rate": 3.790377611279685e-05,
      "loss": 0.567,
      "num_input_tokens_seen": 9841528,
      "step": 15060
    },
    {
      "epoch": 7.895702306079665,
      "grad_norm": 0.6346538066864014,
      "learning_rate": 3.789398132613914e-05,
      "loss": 0.5975,
      "num_input_tokens_seen": 9844984,
      "step": 15065
    },
    {
      "epoch": 7.89832285115304,
      "grad_norm": 0.27674537897109985,
      "learning_rate": 3.788418384220763e-05,
      "loss": 0.4665,
      "num_input_tokens_seen": 9848568,
      "step": 15070
    },
    {
      "epoch": 7.900943396226415,
      "grad_norm": 0.45873868465423584,
      "learning_rate": 3.7874383663051846e-05,
      "loss": 0.2951,
      "num_input_tokens_seen": 9851640,
      "step": 15075
    },
    {
      "epoch": 7.90356394129979,
      "grad_norm": 0.49769580364227295,
      "learning_rate": 3.7864580790721894e-05,
      "loss": 0.3702,
      "num_input_tokens_seen": 9855192,
      "step": 15080
    },
    {
      "epoch": 7.906184486373165,
      "grad_norm": 0.41868025064468384,
      "learning_rate": 3.785477522726839e-05,
      "loss": 0.4686,
      "num_input_tokens_seen": 9858072,
      "step": 15085
    },
    {
      "epoch": 7.908805031446541,
      "grad_norm": 0.4896056652069092,
      "learning_rate": 3.784496697474258e-05,
      "loss": 0.4192,
      "num_input_tokens_seen": 9861784,
      "step": 15090
    },
    {
      "epoch": 7.911425576519916,
      "grad_norm": 0.5831413865089417,
      "learning_rate": 3.783515603519622e-05,
      "loss": 0.5119,
      "num_input_tokens_seen": 9866104,
      "step": 15095
    },
    {
      "epoch": 7.914046121593291,
      "grad_norm": 0.8475841283798218,
      "learning_rate": 3.782534241068164e-05,
      "loss": 0.5906,
      "num_input_tokens_seen": 9869816,
      "step": 15100
    },
    {
      "epoch": 7.916666666666667,
      "grad_norm": 0.37382686138153076,
      "learning_rate": 3.781552610325175e-05,
      "loss": 0.2955,
      "num_input_tokens_seen": 9872504,
      "step": 15105
    },
    {
      "epoch": 7.919287211740042,
      "grad_norm": 0.2995309829711914,
      "learning_rate": 3.780570711496001e-05,
      "loss": 0.4622,
      "num_input_tokens_seen": 9875320,
      "step": 15110
    },
    {
      "epoch": 7.921907756813417,
      "grad_norm": 0.4651682376861572,
      "learning_rate": 3.779588544786044e-05,
      "loss": 0.3706,
      "num_input_tokens_seen": 9878136,
      "step": 15115
    },
    {
      "epoch": 7.9245283018867925,
      "grad_norm": 0.573970377445221,
      "learning_rate": 3.778606110400761e-05,
      "loss": 0.4378,
      "num_input_tokens_seen": 9880792,
      "step": 15120
    },
    {
      "epoch": 7.927148846960168,
      "grad_norm": 0.3101159930229187,
      "learning_rate": 3.7776234085456665e-05,
      "loss": 0.3893,
      "num_input_tokens_seen": 9884312,
      "step": 15125
    },
    {
      "epoch": 7.929769392033543,
      "grad_norm": 0.4368479251861572,
      "learning_rate": 3.776640439426332e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 9887416,
      "step": 15130
    },
    {
      "epoch": 7.932389937106918,
      "grad_norm": 0.38100606203079224,
      "learning_rate": 3.775657203248381e-05,
      "loss": 0.3349,
      "num_input_tokens_seen": 9890200,
      "step": 15135
    },
    {
      "epoch": 7.935010482180293,
      "grad_norm": 0.6743850111961365,
      "learning_rate": 3.774673700217497e-05,
      "loss": 0.3984,
      "num_input_tokens_seen": 9893368,
      "step": 15140
    },
    {
      "epoch": 7.937631027253669,
      "grad_norm": 0.490001916885376,
      "learning_rate": 3.773689930539418e-05,
      "loss": 0.4872,
      "num_input_tokens_seen": 9899224,
      "step": 15145
    },
    {
      "epoch": 7.940251572327044,
      "grad_norm": 0.5535004138946533,
      "learning_rate": 3.7727058944199354e-05,
      "loss": 0.4846,
      "num_input_tokens_seen": 9901944,
      "step": 15150
    },
    {
      "epoch": 7.9428721174004195,
      "grad_norm": 0.26320987939834595,
      "learning_rate": 3.7717215920648994e-05,
      "loss": 0.5831,
      "num_input_tokens_seen": 9905720,
      "step": 15155
    },
    {
      "epoch": 7.945492662473795,
      "grad_norm": 0.44187793135643005,
      "learning_rate": 3.770737023680215e-05,
      "loss": 0.5981,
      "num_input_tokens_seen": 9908440,
      "step": 15160
    },
    {
      "epoch": 7.94811320754717,
      "grad_norm": 0.48611000180244446,
      "learning_rate": 3.7697521894718436e-05,
      "loss": 0.5156,
      "num_input_tokens_seen": 9911256,
      "step": 15165
    },
    {
      "epoch": 7.950733752620545,
      "grad_norm": 0.5395680069923401,
      "learning_rate": 3.7687670896458e-05,
      "loss": 0.4991,
      "num_input_tokens_seen": 9913336,
      "step": 15170
    },
    {
      "epoch": 7.95335429769392,
      "grad_norm": 0.28021490573883057,
      "learning_rate": 3.7677817244081556e-05,
      "loss": 0.3567,
      "num_input_tokens_seen": 9917048,
      "step": 15175
    },
    {
      "epoch": 7.955974842767295,
      "grad_norm": 0.9017734527587891,
      "learning_rate": 3.766796093965038e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 9920952,
      "step": 15180
    },
    {
      "epoch": 7.9585953878406706,
      "grad_norm": 0.4220055341720581,
      "learning_rate": 3.76581019852263e-05,
      "loss": 0.3953,
      "num_input_tokens_seen": 9924376,
      "step": 15185
    },
    {
      "epoch": 7.961215932914046,
      "grad_norm": 0.4223651587963104,
      "learning_rate": 3.7648240382871705e-05,
      "loss": 0.5768,
      "num_input_tokens_seen": 9927160,
      "step": 15190
    },
    {
      "epoch": 7.963836477987421,
      "grad_norm": 0.4499962627887726,
      "learning_rate": 3.763837613464951e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 9930840,
      "step": 15195
    },
    {
      "epoch": 7.966457023060797,
      "grad_norm": 0.3892625570297241,
      "learning_rate": 3.7628509242623215e-05,
      "loss": 0.6375,
      "num_input_tokens_seen": 9934104,
      "step": 15200
    },
    {
      "epoch": 7.969077568134172,
      "grad_norm": 0.5737390518188477,
      "learning_rate": 3.761863970885686e-05,
      "loss": 0.3968,
      "num_input_tokens_seen": 9937016,
      "step": 15205
    },
    {
      "epoch": 7.971698113207547,
      "grad_norm": 1.3153656721115112,
      "learning_rate": 3.760876753541504e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 9939736,
      "step": 15210
    },
    {
      "epoch": 7.9743186582809225,
      "grad_norm": 0.38287612795829773,
      "learning_rate": 3.7598892724362907e-05,
      "loss": 0.501,
      "num_input_tokens_seen": 9942744,
      "step": 15215
    },
    {
      "epoch": 7.976939203354298,
      "grad_norm": 0.868182897567749,
      "learning_rate": 3.758901527776615e-05,
      "loss": 0.577,
      "num_input_tokens_seen": 9946776,
      "step": 15220
    },
    {
      "epoch": 7.979559748427673,
      "grad_norm": 0.5271226763725281,
      "learning_rate": 3.7579135197691015e-05,
      "loss": 0.4716,
      "num_input_tokens_seen": 9949304,
      "step": 15225
    },
    {
      "epoch": 7.982180293501048,
      "grad_norm": 0.3854305148124695,
      "learning_rate": 3.7569252486204306e-05,
      "loss": 0.4608,
      "num_input_tokens_seen": 9953112,
      "step": 15230
    },
    {
      "epoch": 7.984800838574423,
      "grad_norm": 0.457807332277298,
      "learning_rate": 3.755936714537338e-05,
      "loss": 0.476,
      "num_input_tokens_seen": 9956056,
      "step": 15235
    },
    {
      "epoch": 7.987421383647799,
      "grad_norm": 0.6478993892669678,
      "learning_rate": 3.754947917726612e-05,
      "loss": 0.3577,
      "num_input_tokens_seen": 9959416,
      "step": 15240
    },
    {
      "epoch": 7.990041928721174,
      "grad_norm": 0.34971994161605835,
      "learning_rate": 3.7539588583951e-05,
      "loss": 0.3899,
      "num_input_tokens_seen": 9962680,
      "step": 15245
    },
    {
      "epoch": 7.9926624737945495,
      "grad_norm": 0.2748194932937622,
      "learning_rate": 3.752969536749701e-05,
      "loss": 0.5821,
      "num_input_tokens_seen": 9966616,
      "step": 15250
    },
    {
      "epoch": 7.995283018867925,
      "grad_norm": 0.2807104289531708,
      "learning_rate": 3.7519799529973685e-05,
      "loss": 0.4393,
      "num_input_tokens_seen": 9971864,
      "step": 15255
    },
    {
      "epoch": 7.9979035639413,
      "grad_norm": 0.3568064272403717,
      "learning_rate": 3.7509901073451134e-05,
      "loss": 0.7736,
      "num_input_tokens_seen": 9975864,
      "step": 15260
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.4794955849647522,
      "eval_runtime": 9.788,
      "eval_samples_per_second": 86.637,
      "eval_steps_per_second": 21.659,
      "num_input_tokens_seen": 9977520,
      "step": 15264
    },
    {
      "epoch": 8.000524109014675,
      "grad_norm": 0.6135274171829224,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.5284,
      "num_input_tokens_seen": 9978032,
      "step": 15265
    },
    {
      "epoch": 8.00314465408805,
      "grad_norm": 0.4670447111129761,
      "learning_rate": 3.749009631169147e-05,
      "loss": 0.6306,
      "num_input_tokens_seen": 9981616,
      "step": 15270
    },
    {
      "epoch": 8.005765199161425,
      "grad_norm": 0.7170909643173218,
      "learning_rate": 3.74801900105973e-05,
      "loss": 0.3894,
      "num_input_tokens_seen": 9984624,
      "step": 15275
    },
    {
      "epoch": 8.0083857442348,
      "grad_norm": 0.4678058624267578,
      "learning_rate": 3.747028109878975e-05,
      "loss": 0.4267,
      "num_input_tokens_seen": 9987568,
      "step": 15280
    },
    {
      "epoch": 8.011006289308176,
      "grad_norm": 0.4379707872867584,
      "learning_rate": 3.746036957834167e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 9991280,
      "step": 15285
    },
    {
      "epoch": 8.01362683438155,
      "grad_norm": 0.332735151052475,
      "learning_rate": 3.7450455451326415e-05,
      "loss": 0.4315,
      "num_input_tokens_seen": 9994288,
      "step": 15290
    },
    {
      "epoch": 8.016247379454926,
      "grad_norm": 0.4472106099128723,
      "learning_rate": 3.7440538719817935e-05,
      "loss": 0.4106,
      "num_input_tokens_seen": 9997200,
      "step": 15295
    },
    {
      "epoch": 8.018867924528301,
      "grad_norm": 0.4805617928504944,
      "learning_rate": 3.743061938589068e-05,
      "loss": 0.5719,
      "num_input_tokens_seen": 10000176,
      "step": 15300
    },
    {
      "epoch": 8.021488469601676,
      "grad_norm": 0.4099631905555725,
      "learning_rate": 3.742069745161966e-05,
      "loss": 0.4868,
      "num_input_tokens_seen": 10003504,
      "step": 15305
    },
    {
      "epoch": 8.024109014675052,
      "grad_norm": 0.5600143074989319,
      "learning_rate": 3.741077291908045e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 10008688,
      "step": 15310
    },
    {
      "epoch": 8.026729559748428,
      "grad_norm": 0.5036642551422119,
      "learning_rate": 3.740084579034913e-05,
      "loss": 0.5982,
      "num_input_tokens_seen": 10011536,
      "step": 15315
    },
    {
      "epoch": 8.029350104821804,
      "grad_norm": 0.6232349276542664,
      "learning_rate": 3.739091606750234e-05,
      "loss": 0.4401,
      "num_input_tokens_seen": 10014096,
      "step": 15320
    },
    {
      "epoch": 8.031970649895179,
      "grad_norm": 0.27739474177360535,
      "learning_rate": 3.7380983752617275e-05,
      "loss": 0.4328,
      "num_input_tokens_seen": 10017360,
      "step": 15325
    },
    {
      "epoch": 8.034591194968554,
      "grad_norm": 0.3203657269477844,
      "learning_rate": 3.737104884777166e-05,
      "loss": 0.572,
      "num_input_tokens_seen": 10020560,
      "step": 15330
    },
    {
      "epoch": 8.03721174004193,
      "grad_norm": 0.4864533841609955,
      "learning_rate": 3.736111135504376e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 10024176,
      "step": 15335
    },
    {
      "epoch": 8.039832285115304,
      "grad_norm": 0.34468016028404236,
      "learning_rate": 3.735117127651239e-05,
      "loss": 0.3244,
      "num_input_tokens_seen": 10028336,
      "step": 15340
    },
    {
      "epoch": 8.04245283018868,
      "grad_norm": 0.5052928328514099,
      "learning_rate": 3.734122861425689e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 10031376,
      "step": 15345
    },
    {
      "epoch": 8.045073375262055,
      "grad_norm": 0.6189833879470825,
      "learning_rate": 3.7331283370357173e-05,
      "loss": 0.4527,
      "num_input_tokens_seen": 10033904,
      "step": 15350
    },
    {
      "epoch": 8.04769392033543,
      "grad_norm": 0.3811035752296448,
      "learning_rate": 3.732133554689365e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 10036720,
      "step": 15355
    },
    {
      "epoch": 8.050314465408805,
      "grad_norm": 0.4869700074195862,
      "learning_rate": 3.73113851459473e-05,
      "loss": 0.5027,
      "num_input_tokens_seen": 10040016,
      "step": 15360
    },
    {
      "epoch": 8.05293501048218,
      "grad_norm": 0.3229936957359314,
      "learning_rate": 3.730143216959964e-05,
      "loss": 0.5194,
      "num_input_tokens_seen": 10043088,
      "step": 15365
    },
    {
      "epoch": 8.055555555555555,
      "grad_norm": 0.4713076055049896,
      "learning_rate": 3.729147661993271e-05,
      "loss": 0.5441,
      "num_input_tokens_seen": 10046160,
      "step": 15370
    },
    {
      "epoch": 8.05817610062893,
      "grad_norm": 0.39128273725509644,
      "learning_rate": 3.728151849902909e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 10048592,
      "step": 15375
    },
    {
      "epoch": 8.060796645702306,
      "grad_norm": 0.20060202479362488,
      "learning_rate": 3.727155780897193e-05,
      "loss": 0.4872,
      "num_input_tokens_seen": 10051536,
      "step": 15380
    },
    {
      "epoch": 8.06341719077568,
      "grad_norm": 0.3494293689727783,
      "learning_rate": 3.7261594551844865e-05,
      "loss": 0.5232,
      "num_input_tokens_seen": 10055632,
      "step": 15385
    },
    {
      "epoch": 8.066037735849056,
      "grad_norm": 0.361960232257843,
      "learning_rate": 3.7251628729732116e-05,
      "loss": 0.5122,
      "num_input_tokens_seen": 10060528,
      "step": 15390
    },
    {
      "epoch": 8.068658280922431,
      "grad_norm": 0.5712471604347229,
      "learning_rate": 3.724166034471841e-05,
      "loss": 0.3402,
      "num_input_tokens_seen": 10063664,
      "step": 15395
    },
    {
      "epoch": 8.071278825995806,
      "grad_norm": 0.911848247051239,
      "learning_rate": 3.723168939888901e-05,
      "loss": 0.4766,
      "num_input_tokens_seen": 10066288,
      "step": 15400
    },
    {
      "epoch": 8.073899371069182,
      "grad_norm": 0.9320734143257141,
      "learning_rate": 3.7221715894329747e-05,
      "loss": 0.5961,
      "num_input_tokens_seen": 10070064,
      "step": 15405
    },
    {
      "epoch": 8.076519916142558,
      "grad_norm": 0.4564517140388489,
      "learning_rate": 3.7211739833126945e-05,
      "loss": 0.3882,
      "num_input_tokens_seen": 10074128,
      "step": 15410
    },
    {
      "epoch": 8.079140461215934,
      "grad_norm": 0.3157099485397339,
      "learning_rate": 3.720176121736748e-05,
      "loss": 0.294,
      "num_input_tokens_seen": 10077936,
      "step": 15415
    },
    {
      "epoch": 8.081761006289309,
      "grad_norm": 0.44494307041168213,
      "learning_rate": 3.719178004913879e-05,
      "loss": 0.4848,
      "num_input_tokens_seen": 10080976,
      "step": 15420
    },
    {
      "epoch": 8.084381551362684,
      "grad_norm": 0.5137907862663269,
      "learning_rate": 3.7181796330528796e-05,
      "loss": 0.4642,
      "num_input_tokens_seen": 10083984,
      "step": 15425
    },
    {
      "epoch": 8.08700209643606,
      "grad_norm": 0.5621265172958374,
      "learning_rate": 3.717181006362599e-05,
      "loss": 0.5908,
      "num_input_tokens_seen": 10086800,
      "step": 15430
    },
    {
      "epoch": 8.089622641509434,
      "grad_norm": 0.3404883146286011,
      "learning_rate": 3.7161821250519384e-05,
      "loss": 0.4022,
      "num_input_tokens_seen": 10090544,
      "step": 15435
    },
    {
      "epoch": 8.09224318658281,
      "grad_norm": 1.2212687730789185,
      "learning_rate": 3.7151829893298515e-05,
      "loss": 0.4137,
      "num_input_tokens_seen": 10093360,
      "step": 15440
    },
    {
      "epoch": 8.094863731656185,
      "grad_norm": 0.42073166370391846,
      "learning_rate": 3.714183599405347e-05,
      "loss": 0.5894,
      "num_input_tokens_seen": 10097520,
      "step": 15445
    },
    {
      "epoch": 8.09748427672956,
      "grad_norm": 0.37637466192245483,
      "learning_rate": 3.713183955487487e-05,
      "loss": 0.402,
      "num_input_tokens_seen": 10100240,
      "step": 15450
    },
    {
      "epoch": 8.100104821802935,
      "grad_norm": 0.4118175208568573,
      "learning_rate": 3.7121840577853815e-05,
      "loss": 0.4337,
      "num_input_tokens_seen": 10103280,
      "step": 15455
    },
    {
      "epoch": 8.10272536687631,
      "grad_norm": 0.28938204050064087,
      "learning_rate": 3.7111839065082027e-05,
      "loss": 0.4876,
      "num_input_tokens_seen": 10106288,
      "step": 15460
    },
    {
      "epoch": 8.105345911949685,
      "grad_norm": 0.39624589681625366,
      "learning_rate": 3.710183501865168e-05,
      "loss": 0.3974,
      "num_input_tokens_seen": 10110064,
      "step": 15465
    },
    {
      "epoch": 8.10796645702306,
      "grad_norm": 0.32770758867263794,
      "learning_rate": 3.7091828440655514e-05,
      "loss": 0.364,
      "num_input_tokens_seen": 10113360,
      "step": 15470
    },
    {
      "epoch": 8.110587002096436,
      "grad_norm": 0.46231555938720703,
      "learning_rate": 3.708181933318678e-05,
      "loss": 0.5151,
      "num_input_tokens_seen": 10119760,
      "step": 15475
    },
    {
      "epoch": 8.11320754716981,
      "grad_norm": 0.33677393198013306,
      "learning_rate": 3.707180769833929e-05,
      "loss": 0.4214,
      "num_input_tokens_seen": 10126864,
      "step": 15480
    },
    {
      "epoch": 8.115828092243186,
      "grad_norm": 0.3855981230735779,
      "learning_rate": 3.706179353820734e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 10129776,
      "step": 15485
    },
    {
      "epoch": 8.118448637316561,
      "grad_norm": 0.5050025582313538,
      "learning_rate": 3.7051776854885794e-05,
      "loss": 0.5228,
      "num_input_tokens_seen": 10132816,
      "step": 15490
    },
    {
      "epoch": 8.121069182389936,
      "grad_norm": 0.47342991828918457,
      "learning_rate": 3.7041757650470025e-05,
      "loss": 0.3293,
      "num_input_tokens_seen": 10136112,
      "step": 15495
    },
    {
      "epoch": 8.123689727463312,
      "grad_norm": 0.2568972408771515,
      "learning_rate": 3.703173592705594e-05,
      "loss": 0.4422,
      "num_input_tokens_seen": 10138896,
      "step": 15500
    },
    {
      "epoch": 8.126310272536688,
      "grad_norm": 0.37076321244239807,
      "learning_rate": 3.702171168673996e-05,
      "loss": 0.4496,
      "num_input_tokens_seen": 10141456,
      "step": 15505
    },
    {
      "epoch": 8.128930817610064,
      "grad_norm": 0.4014833867549896,
      "learning_rate": 3.701168493161905e-05,
      "loss": 0.5275,
      "num_input_tokens_seen": 10144944,
      "step": 15510
    },
    {
      "epoch": 8.131551362683439,
      "grad_norm": 0.361562043428421,
      "learning_rate": 3.7001655663790676e-05,
      "loss": 0.3874,
      "num_input_tokens_seen": 10148016,
      "step": 15515
    },
    {
      "epoch": 8.134171907756814,
      "grad_norm": 0.36578965187072754,
      "learning_rate": 3.699162388535288e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 10152048,
      "step": 15520
    },
    {
      "epoch": 8.13679245283019,
      "grad_norm": 0.3743579387664795,
      "learning_rate": 3.698158959840416e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 10155600,
      "step": 15525
    },
    {
      "epoch": 8.139412997903564,
      "grad_norm": 0.6600983738899231,
      "learning_rate": 3.6971552805043585e-05,
      "loss": 0.481,
      "num_input_tokens_seen": 10158160,
      "step": 15530
    },
    {
      "epoch": 8.14203354297694,
      "grad_norm": 0.31654414534568787,
      "learning_rate": 3.6961513507370754e-05,
      "loss": 0.4776,
      "num_input_tokens_seen": 10161296,
      "step": 15535
    },
    {
      "epoch": 8.144654088050315,
      "grad_norm": 0.26715564727783203,
      "learning_rate": 3.695147170748576e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 10164112,
      "step": 15540
    },
    {
      "epoch": 8.14727463312369,
      "grad_norm": 0.5374179482460022,
      "learning_rate": 3.694142740748923e-05,
      "loss": 0.4511,
      "num_input_tokens_seen": 10167248,
      "step": 15545
    },
    {
      "epoch": 8.149895178197065,
      "grad_norm": 0.38694193959236145,
      "learning_rate": 3.693138060948232e-05,
      "loss": 0.3593,
      "num_input_tokens_seen": 10171088,
      "step": 15550
    },
    {
      "epoch": 8.15251572327044,
      "grad_norm": 0.4976344406604767,
      "learning_rate": 3.692133131556671e-05,
      "loss": 0.4855,
      "num_input_tokens_seen": 10174192,
      "step": 15555
    },
    {
      "epoch": 8.155136268343815,
      "grad_norm": 0.44936662912368774,
      "learning_rate": 3.6911279527844596e-05,
      "loss": 0.4952,
      "num_input_tokens_seen": 10177264,
      "step": 15560
    },
    {
      "epoch": 8.15775681341719,
      "grad_norm": 0.7270626425743103,
      "learning_rate": 3.69012252484187e-05,
      "loss": 0.5671,
      "num_input_tokens_seen": 10179824,
      "step": 15565
    },
    {
      "epoch": 8.160377358490566,
      "grad_norm": 0.34481698274612427,
      "learning_rate": 3.6891168479392245e-05,
      "loss": 0.4235,
      "num_input_tokens_seen": 10183024,
      "step": 15570
    },
    {
      "epoch": 8.16299790356394,
      "grad_norm": 0.32266780734062195,
      "learning_rate": 3.688110922286901e-05,
      "loss": 0.4315,
      "num_input_tokens_seen": 10186832,
      "step": 15575
    },
    {
      "epoch": 8.165618448637316,
      "grad_norm": 0.4386199414730072,
      "learning_rate": 3.687104748095328e-05,
      "loss": 0.5062,
      "num_input_tokens_seen": 10189360,
      "step": 15580
    },
    {
      "epoch": 8.168238993710691,
      "grad_norm": 0.3764984607696533,
      "learning_rate": 3.686098325574984e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 10193360,
      "step": 15585
    },
    {
      "epoch": 8.170859538784066,
      "grad_norm": 0.5905779600143433,
      "learning_rate": 3.685091654936403e-05,
      "loss": 0.4624,
      "num_input_tokens_seen": 10197136,
      "step": 15590
    },
    {
      "epoch": 8.173480083857442,
      "grad_norm": 0.4298696517944336,
      "learning_rate": 3.6840847363901675e-05,
      "loss": 0.4554,
      "num_input_tokens_seen": 10200592,
      "step": 15595
    },
    {
      "epoch": 8.176100628930818,
      "grad_norm": 0.5408369898796082,
      "learning_rate": 3.683077570146914e-05,
      "loss": 0.4469,
      "num_input_tokens_seen": 10203088,
      "step": 15600
    },
    {
      "epoch": 8.178721174004194,
      "grad_norm": 0.5322877764701843,
      "learning_rate": 3.68207015641733e-05,
      "loss": 0.5987,
      "num_input_tokens_seen": 10205840,
      "step": 15605
    },
    {
      "epoch": 8.181341719077569,
      "grad_norm": 0.35895198583602905,
      "learning_rate": 3.681062495412155e-05,
      "loss": 0.3576,
      "num_input_tokens_seen": 10209040,
      "step": 15610
    },
    {
      "epoch": 8.183962264150944,
      "grad_norm": 0.43652108311653137,
      "learning_rate": 3.68005458734218e-05,
      "loss": 0.4306,
      "num_input_tokens_seen": 10212048,
      "step": 15615
    },
    {
      "epoch": 8.18658280922432,
      "grad_norm": 0.44654932618141174,
      "learning_rate": 3.679046432418248e-05,
      "loss": 0.4978,
      "num_input_tokens_seen": 10214896,
      "step": 15620
    },
    {
      "epoch": 8.189203354297694,
      "grad_norm": 0.47307470440864563,
      "learning_rate": 3.6780380308512543e-05,
      "loss": 0.4709,
      "num_input_tokens_seen": 10217872,
      "step": 15625
    },
    {
      "epoch": 8.19182389937107,
      "grad_norm": 0.5532592535018921,
      "learning_rate": 3.677029382852144e-05,
      "loss": 0.4567,
      "num_input_tokens_seen": 10221488,
      "step": 15630
    },
    {
      "epoch": 8.194444444444445,
      "grad_norm": 0.48171916604042053,
      "learning_rate": 3.676020488631915e-05,
      "loss": 0.3697,
      "num_input_tokens_seen": 10224048,
      "step": 15635
    },
    {
      "epoch": 8.19706498951782,
      "grad_norm": 0.3098432123661041,
      "learning_rate": 3.675011348401617e-05,
      "loss": 0.4057,
      "num_input_tokens_seen": 10228368,
      "step": 15640
    },
    {
      "epoch": 8.199685534591195,
      "grad_norm": 0.3692972958087921,
      "learning_rate": 3.6740019623723494e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 10231728,
      "step": 15645
    },
    {
      "epoch": 8.20230607966457,
      "grad_norm": 0.42722123861312866,
      "learning_rate": 3.672992330755265e-05,
      "loss": 0.429,
      "num_input_tokens_seen": 10234960,
      "step": 15650
    },
    {
      "epoch": 8.204926624737945,
      "grad_norm": 0.3041529059410095,
      "learning_rate": 3.671982453761568e-05,
      "loss": 0.4263,
      "num_input_tokens_seen": 10238448,
      "step": 15655
    },
    {
      "epoch": 8.20754716981132,
      "grad_norm": 0.5763108730316162,
      "learning_rate": 3.6709723316025114e-05,
      "loss": 0.3703,
      "num_input_tokens_seen": 10240912,
      "step": 15660
    },
    {
      "epoch": 8.210167714884696,
      "grad_norm": 0.31824034452438354,
      "learning_rate": 3.669961964489403e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 10244816,
      "step": 15665
    },
    {
      "epoch": 8.21278825995807,
      "grad_norm": 0.4804253578186035,
      "learning_rate": 3.668951352633599e-05,
      "loss": 0.4849,
      "num_input_tokens_seen": 10247728,
      "step": 15670
    },
    {
      "epoch": 8.215408805031446,
      "grad_norm": 0.33307084441185,
      "learning_rate": 3.6679404962465084e-05,
      "loss": 0.4054,
      "num_input_tokens_seen": 10251248,
      "step": 15675
    },
    {
      "epoch": 8.218029350104821,
      "grad_norm": 0.43696078658103943,
      "learning_rate": 3.66692939553959e-05,
      "loss": 0.4692,
      "num_input_tokens_seen": 10254160,
      "step": 15680
    },
    {
      "epoch": 8.220649895178196,
      "grad_norm": 0.39868998527526855,
      "learning_rate": 3.6659180507243554e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 10256944,
      "step": 15685
    },
    {
      "epoch": 8.223270440251572,
      "grad_norm": 0.4784729480743408,
      "learning_rate": 3.664906462012366e-05,
      "loss": 0.5118,
      "num_input_tokens_seen": 10259888,
      "step": 15690
    },
    {
      "epoch": 8.225890985324948,
      "grad_norm": 0.4353800117969513,
      "learning_rate": 3.663894629615234e-05,
      "loss": 0.5721,
      "num_input_tokens_seen": 10262640,
      "step": 15695
    },
    {
      "epoch": 8.228511530398324,
      "grad_norm": 0.5069556832313538,
      "learning_rate": 3.6628825537446246e-05,
      "loss": 0.4933,
      "num_input_tokens_seen": 10265424,
      "step": 15700
    },
    {
      "epoch": 8.231132075471699,
      "grad_norm": 0.3997410833835602,
      "learning_rate": 3.6618702346122525e-05,
      "loss": 0.4959,
      "num_input_tokens_seen": 10267888,
      "step": 15705
    },
    {
      "epoch": 8.233752620545074,
      "grad_norm": 0.20959018170833588,
      "learning_rate": 3.6608576724298817e-05,
      "loss": 0.3222,
      "num_input_tokens_seen": 10270992,
      "step": 15710
    },
    {
      "epoch": 8.23637316561845,
      "grad_norm": 0.6665362119674683,
      "learning_rate": 3.65984486740933e-05,
      "loss": 0.5017,
      "num_input_tokens_seen": 10274128,
      "step": 15715
    },
    {
      "epoch": 8.238993710691824,
      "grad_norm": 0.6283776760101318,
      "learning_rate": 3.6588318197624644e-05,
      "loss": 0.4082,
      "num_input_tokens_seen": 10278128,
      "step": 15720
    },
    {
      "epoch": 8.2416142557652,
      "grad_norm": 0.4213528037071228,
      "learning_rate": 3.6578185297012015e-05,
      "loss": 0.4335,
      "num_input_tokens_seen": 10280720,
      "step": 15725
    },
    {
      "epoch": 8.244234800838575,
      "grad_norm": 0.3310258090496063,
      "learning_rate": 3.656804997437512e-05,
      "loss": 0.4287,
      "num_input_tokens_seen": 10283280,
      "step": 15730
    },
    {
      "epoch": 8.24685534591195,
      "grad_norm": 0.333676815032959,
      "learning_rate": 3.6557912231834155e-05,
      "loss": 0.4848,
      "num_input_tokens_seen": 10287600,
      "step": 15735
    },
    {
      "epoch": 8.249475890985325,
      "grad_norm": 0.6031058430671692,
      "learning_rate": 3.65477720715098e-05,
      "loss": 0.4198,
      "num_input_tokens_seen": 10289808,
      "step": 15740
    },
    {
      "epoch": 8.2520964360587,
      "grad_norm": 0.7495798468589783,
      "learning_rate": 3.653762949552328e-05,
      "loss": 0.498,
      "num_input_tokens_seen": 10292656,
      "step": 15745
    },
    {
      "epoch": 8.254716981132075,
      "grad_norm": 0.3746139705181122,
      "learning_rate": 3.6527484505996295e-05,
      "loss": 0.4042,
      "num_input_tokens_seen": 10295664,
      "step": 15750
    },
    {
      "epoch": 8.25733752620545,
      "grad_norm": 0.6545858979225159,
      "learning_rate": 3.651733710505106e-05,
      "loss": 0.4874,
      "num_input_tokens_seen": 10299248,
      "step": 15755
    },
    {
      "epoch": 8.259958071278826,
      "grad_norm": 0.5375692844390869,
      "learning_rate": 3.6507187294810305e-05,
      "loss": 0.4879,
      "num_input_tokens_seen": 10302832,
      "step": 15760
    },
    {
      "epoch": 8.2625786163522,
      "grad_norm": 0.32085713744163513,
      "learning_rate": 3.649703507739724e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 10307888,
      "step": 15765
    },
    {
      "epoch": 8.265199161425576,
      "grad_norm": 0.556512713432312,
      "learning_rate": 3.6486880454935606e-05,
      "loss": 0.4326,
      "num_input_tokens_seen": 10311248,
      "step": 15770
    },
    {
      "epoch": 8.267819706498951,
      "grad_norm": 0.5178356170654297,
      "learning_rate": 3.647672342954963e-05,
      "loss": 0.4868,
      "num_input_tokens_seen": 10314320,
      "step": 15775
    },
    {
      "epoch": 8.270440251572326,
      "grad_norm": 0.457839697599411,
      "learning_rate": 3.646656400336405e-05,
      "loss": 0.492,
      "num_input_tokens_seen": 10317936,
      "step": 15780
    },
    {
      "epoch": 8.273060796645701,
      "grad_norm": 0.42022180557250977,
      "learning_rate": 3.6456402178504094e-05,
      "loss": 0.529,
      "num_input_tokens_seen": 10321136,
      "step": 15785
    },
    {
      "epoch": 8.275681341719078,
      "grad_norm": 0.4084852933883667,
      "learning_rate": 3.6446237957095505e-05,
      "loss": 0.527,
      "num_input_tokens_seen": 10323760,
      "step": 15790
    },
    {
      "epoch": 8.278301886792454,
      "grad_norm": 1.1655155420303345,
      "learning_rate": 3.643607134126452e-05,
      "loss": 0.4863,
      "num_input_tokens_seen": 10327088,
      "step": 15795
    },
    {
      "epoch": 8.280922431865829,
      "grad_norm": 0.49590829014778137,
      "learning_rate": 3.642590233313789e-05,
      "loss": 0.5773,
      "num_input_tokens_seen": 10330448,
      "step": 15800
    },
    {
      "epoch": 8.283542976939204,
      "grad_norm": 0.693443238735199,
      "learning_rate": 3.6415730934842827e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 10334672,
      "step": 15805
    },
    {
      "epoch": 8.286163522012579,
      "grad_norm": 0.6362709403038025,
      "learning_rate": 3.64055571485071e-05,
      "loss": 0.5033,
      "num_input_tokens_seen": 10337904,
      "step": 15810
    },
    {
      "epoch": 8.288784067085954,
      "grad_norm": 0.5411792993545532,
      "learning_rate": 3.639538097625894e-05,
      "loss": 0.4197,
      "num_input_tokens_seen": 10341488,
      "step": 15815
    },
    {
      "epoch": 8.29140461215933,
      "grad_norm": 0.5469892621040344,
      "learning_rate": 3.638520242022708e-05,
      "loss": 0.4571,
      "num_input_tokens_seen": 10344944,
      "step": 15820
    },
    {
      "epoch": 8.294025157232705,
      "grad_norm": 0.7033849954605103,
      "learning_rate": 3.637502148254077e-05,
      "loss": 0.5308,
      "num_input_tokens_seen": 10347248,
      "step": 15825
    },
    {
      "epoch": 8.29664570230608,
      "grad_norm": 0.4569484293460846,
      "learning_rate": 3.636483816532974e-05,
      "loss": 0.5766,
      "num_input_tokens_seen": 10349712,
      "step": 15830
    },
    {
      "epoch": 8.299266247379455,
      "grad_norm": 0.4540281891822815,
      "learning_rate": 3.6354652470724217e-05,
      "loss": 0.4523,
      "num_input_tokens_seen": 10353936,
      "step": 15835
    },
    {
      "epoch": 8.30188679245283,
      "grad_norm": 1.2492380142211914,
      "learning_rate": 3.634446440085494e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 10356432,
      "step": 15840
    },
    {
      "epoch": 8.304507337526205,
      "grad_norm": 0.5387322902679443,
      "learning_rate": 3.633427395785313e-05,
      "loss": 0.7048,
      "num_input_tokens_seen": 10360624,
      "step": 15845
    },
    {
      "epoch": 8.30712788259958,
      "grad_norm": 0.5007079839706421,
      "learning_rate": 3.6324081143850516e-05,
      "loss": 0.4771,
      "num_input_tokens_seen": 10364432,
      "step": 15850
    },
    {
      "epoch": 8.309748427672956,
      "grad_norm": 0.4333820343017578,
      "learning_rate": 3.6313885960979326e-05,
      "loss": 0.4325,
      "num_input_tokens_seen": 10367184,
      "step": 15855
    },
    {
      "epoch": 8.31236897274633,
      "grad_norm": 0.47406262159347534,
      "learning_rate": 3.630368841137226e-05,
      "loss": 0.4352,
      "num_input_tokens_seen": 10370192,
      "step": 15860
    },
    {
      "epoch": 8.314989517819706,
      "grad_norm": 0.39439672231674194,
      "learning_rate": 3.6293488497162554e-05,
      "loss": 0.54,
      "num_input_tokens_seen": 10373744,
      "step": 15865
    },
    {
      "epoch": 8.317610062893081,
      "grad_norm": 0.42424190044403076,
      "learning_rate": 3.628328622048388e-05,
      "loss": 0.4972,
      "num_input_tokens_seen": 10376400,
      "step": 15870
    },
    {
      "epoch": 8.320230607966456,
      "grad_norm": 0.3762827515602112,
      "learning_rate": 3.627308158347046e-05,
      "loss": 0.3594,
      "num_input_tokens_seen": 10379600,
      "step": 15875
    },
    {
      "epoch": 8.322851153039831,
      "grad_norm": 0.5040093660354614,
      "learning_rate": 3.626287458825699e-05,
      "loss": 0.4408,
      "num_input_tokens_seen": 10383088,
      "step": 15880
    },
    {
      "epoch": 8.325471698113208,
      "grad_norm": 0.5219465494155884,
      "learning_rate": 3.625266523697864e-05,
      "loss": 0.4736,
      "num_input_tokens_seen": 10386288,
      "step": 15885
    },
    {
      "epoch": 8.328092243186584,
      "grad_norm": 0.3564211130142212,
      "learning_rate": 3.62424535317711e-05,
      "loss": 0.6154,
      "num_input_tokens_seen": 10389744,
      "step": 15890
    },
    {
      "epoch": 8.330712788259959,
      "grad_norm": 0.40691259503364563,
      "learning_rate": 3.623223947477055e-05,
      "loss": 0.383,
      "num_input_tokens_seen": 10392880,
      "step": 15895
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 0.612611711025238,
      "learning_rate": 3.6222023068113635e-05,
      "loss": 0.5657,
      "num_input_tokens_seen": 10397168,
      "step": 15900
    },
    {
      "epoch": 8.335953878406709,
      "grad_norm": 0.661285936832428,
      "learning_rate": 3.621180431393754e-05,
      "loss": 0.4771,
      "num_input_tokens_seen": 10400784,
      "step": 15905
    },
    {
      "epoch": 8.338574423480084,
      "grad_norm": 0.4307464063167572,
      "learning_rate": 3.6201583214379874e-05,
      "loss": 0.2917,
      "num_input_tokens_seen": 10403568,
      "step": 15910
    },
    {
      "epoch": 8.34119496855346,
      "grad_norm": 0.5555877089500427,
      "learning_rate": 3.61913597715788e-05,
      "loss": 0.4411,
      "num_input_tokens_seen": 10406800,
      "step": 15915
    },
    {
      "epoch": 8.343815513626835,
      "grad_norm": 0.2986058294773102,
      "learning_rate": 3.6181133987672944e-05,
      "loss": 0.3807,
      "num_input_tokens_seen": 10409744,
      "step": 15920
    },
    {
      "epoch": 8.34643605870021,
      "grad_norm": 0.5802407264709473,
      "learning_rate": 3.617090586480142e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 10412880,
      "step": 15925
    },
    {
      "epoch": 8.349056603773585,
      "grad_norm": 0.6068930625915527,
      "learning_rate": 3.6160675405103826e-05,
      "loss": 0.4764,
      "num_input_tokens_seen": 10416432,
      "step": 15930
    },
    {
      "epoch": 8.35167714884696,
      "grad_norm": 0.7438477873802185,
      "learning_rate": 3.615044261072027e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 10419824,
      "step": 15935
    },
    {
      "epoch": 8.354297693920335,
      "grad_norm": 0.3456224501132965,
      "learning_rate": 3.614020748379133e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 10422800,
      "step": 15940
    },
    {
      "epoch": 8.35691823899371,
      "grad_norm": 0.5111890435218811,
      "learning_rate": 3.6129970026458076e-05,
      "loss": 0.3893,
      "num_input_tokens_seen": 10425808,
      "step": 15945
    },
    {
      "epoch": 8.359538784067086,
      "grad_norm": 0.3447943329811096,
      "learning_rate": 3.611973024086207e-05,
      "loss": 0.453,
      "num_input_tokens_seen": 10428944,
      "step": 15950
    },
    {
      "epoch": 8.36215932914046,
      "grad_norm": 0.4088447093963623,
      "learning_rate": 3.610948812914536e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 10433648,
      "step": 15955
    },
    {
      "epoch": 8.364779874213836,
      "grad_norm": 0.2926360070705414,
      "learning_rate": 3.609924369345048e-05,
      "loss": 0.432,
      "num_input_tokens_seen": 10436784,
      "step": 15960
    },
    {
      "epoch": 8.367400419287211,
      "grad_norm": 0.5657051801681519,
      "learning_rate": 3.6088996935920444e-05,
      "loss": 0.4411,
      "num_input_tokens_seen": 10439280,
      "step": 15965
    },
    {
      "epoch": 8.370020964360586,
      "grad_norm": 0.38394659757614136,
      "learning_rate": 3.607874785869876e-05,
      "loss": 0.3695,
      "num_input_tokens_seen": 10441744,
      "step": 15970
    },
    {
      "epoch": 8.372641509433961,
      "grad_norm": 0.3908923864364624,
      "learning_rate": 3.6068496463929416e-05,
      "loss": 0.4308,
      "num_input_tokens_seen": 10445424,
      "step": 15975
    },
    {
      "epoch": 8.375262054507338,
      "grad_norm": 0.550102174282074,
      "learning_rate": 3.605824275375689e-05,
      "loss": 0.649,
      "num_input_tokens_seen": 10448688,
      "step": 15980
    },
    {
      "epoch": 8.377882599580714,
      "grad_norm": 0.41250747442245483,
      "learning_rate": 3.604798673032613e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 10453040,
      "step": 15985
    },
    {
      "epoch": 8.380503144654089,
      "grad_norm": 0.444715678691864,
      "learning_rate": 3.60377283957826e-05,
      "loss": 0.5155,
      "num_input_tokens_seen": 10457072,
      "step": 15990
    },
    {
      "epoch": 8.383123689727464,
      "grad_norm": 1.4167678356170654,
      "learning_rate": 3.6027467752272203e-05,
      "loss": 0.3312,
      "num_input_tokens_seen": 10459280,
      "step": 15995
    },
    {
      "epoch": 8.385744234800839,
      "grad_norm": 0.7528859376907349,
      "learning_rate": 3.601720480194137e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 10462192,
      "step": 16000
    },
    {
      "epoch": 8.388364779874214,
      "grad_norm": 0.3820917308330536,
      "learning_rate": 3.600693954693698e-05,
      "loss": 0.4586,
      "num_input_tokens_seen": 10466672,
      "step": 16005
    },
    {
      "epoch": 8.39098532494759,
      "grad_norm": 0.33763015270233154,
      "learning_rate": 3.5996671989406414e-05,
      "loss": 0.45,
      "num_input_tokens_seen": 10469776,
      "step": 16010
    },
    {
      "epoch": 8.393605870020965,
      "grad_norm": 0.7951868176460266,
      "learning_rate": 3.598640213149752e-05,
      "loss": 0.3521,
      "num_input_tokens_seen": 10472848,
      "step": 16015
    },
    {
      "epoch": 8.39622641509434,
      "grad_norm": 1.3898845911026,
      "learning_rate": 3.597612997535864e-05,
      "loss": 0.5325,
      "num_input_tokens_seen": 10476368,
      "step": 16020
    },
    {
      "epoch": 8.398846960167715,
      "grad_norm": 0.4373679757118225,
      "learning_rate": 3.596585552313859e-05,
      "loss": 0.4735,
      "num_input_tokens_seen": 10479472,
      "step": 16025
    },
    {
      "epoch": 8.40146750524109,
      "grad_norm": 0.5357080101966858,
      "learning_rate": 3.595557877698666e-05,
      "loss": 0.5242,
      "num_input_tokens_seen": 10482000,
      "step": 16030
    },
    {
      "epoch": 8.404088050314465,
      "grad_norm": 0.5818850994110107,
      "learning_rate": 3.594529973905266e-05,
      "loss": 0.4719,
      "num_input_tokens_seen": 10484592,
      "step": 16035
    },
    {
      "epoch": 8.40670859538784,
      "grad_norm": 1.018703818321228,
      "learning_rate": 3.593501841148681e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 10487216,
      "step": 16040
    },
    {
      "epoch": 8.409329140461216,
      "grad_norm": 0.4157094657421112,
      "learning_rate": 3.5924734796439864e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 10490512,
      "step": 16045
    },
    {
      "epoch": 8.41194968553459,
      "grad_norm": 0.450870543718338,
      "learning_rate": 3.5914448896063044e-05,
      "loss": 0.3593,
      "num_input_tokens_seen": 10493552,
      "step": 16050
    },
    {
      "epoch": 8.414570230607966,
      "grad_norm": 0.4976455867290497,
      "learning_rate": 3.590416071250802e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 10497264,
      "step": 16055
    },
    {
      "epoch": 8.417190775681341,
      "grad_norm": 0.4270634055137634,
      "learning_rate": 3.589387024792699e-05,
      "loss": 0.3558,
      "num_input_tokens_seen": 10499984,
      "step": 16060
    },
    {
      "epoch": 8.419811320754716,
      "grad_norm": 0.5055228471755981,
      "learning_rate": 3.5883577504472576e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 10503248,
      "step": 16065
    },
    {
      "epoch": 8.422431865828091,
      "grad_norm": 0.35050642490386963,
      "learning_rate": 3.587328248429792e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 10507088,
      "step": 16070
    },
    {
      "epoch": 8.425052410901468,
      "grad_norm": 0.35215330123901367,
      "learning_rate": 3.586298518955661e-05,
      "loss": 0.4201,
      "num_input_tokens_seen": 10510416,
      "step": 16075
    },
    {
      "epoch": 8.427672955974844,
      "grad_norm": 0.4795396625995636,
      "learning_rate": 3.5852685622402746e-05,
      "loss": 0.5199,
      "num_input_tokens_seen": 10514384,
      "step": 16080
    },
    {
      "epoch": 8.430293501048219,
      "grad_norm": 0.328095942735672,
      "learning_rate": 3.584238378499085e-05,
      "loss": 0.5473,
      "num_input_tokens_seen": 10517872,
      "step": 16085
    },
    {
      "epoch": 8.432914046121594,
      "grad_norm": 0.46407970786094666,
      "learning_rate": 3.5832079679475974e-05,
      "loss": 0.4728,
      "num_input_tokens_seen": 10521200,
      "step": 16090
    },
    {
      "epoch": 8.435534591194969,
      "grad_norm": 0.3755764961242676,
      "learning_rate": 3.582177330801359e-05,
      "loss": 0.4417,
      "num_input_tokens_seen": 10523856,
      "step": 16095
    },
    {
      "epoch": 8.438155136268344,
      "grad_norm": 0.3183318078517914,
      "learning_rate": 3.58114646727597e-05,
      "loss": 0.4958,
      "num_input_tokens_seen": 10527728,
      "step": 16100
    },
    {
      "epoch": 8.44077568134172,
      "grad_norm": 0.45954760909080505,
      "learning_rate": 3.580115377587075e-05,
      "loss": 0.4519,
      "num_input_tokens_seen": 10531536,
      "step": 16105
    },
    {
      "epoch": 8.443396226415095,
      "grad_norm": 0.3393147587776184,
      "learning_rate": 3.5790840619503654e-05,
      "loss": 0.3953,
      "num_input_tokens_seen": 10534352,
      "step": 16110
    },
    {
      "epoch": 8.44601677148847,
      "grad_norm": 0.36246252059936523,
      "learning_rate": 3.57805252058158e-05,
      "loss": 0.461,
      "num_input_tokens_seen": 10537552,
      "step": 16115
    },
    {
      "epoch": 8.448637316561845,
      "grad_norm": 0.6273065209388733,
      "learning_rate": 3.577020753696507e-05,
      "loss": 0.4542,
      "num_input_tokens_seen": 10539952,
      "step": 16120
    },
    {
      "epoch": 8.45125786163522,
      "grad_norm": 0.4504236578941345,
      "learning_rate": 3.57598876151098e-05,
      "loss": 0.4316,
      "num_input_tokens_seen": 10543408,
      "step": 16125
    },
    {
      "epoch": 8.453878406708595,
      "grad_norm": 0.28706255555152893,
      "learning_rate": 3.5749565442408796e-05,
      "loss": 0.3573,
      "num_input_tokens_seen": 10546352,
      "step": 16130
    },
    {
      "epoch": 8.45649895178197,
      "grad_norm": 0.31609705090522766,
      "learning_rate": 3.573924102102133e-05,
      "loss": 0.4014,
      "num_input_tokens_seen": 10550352,
      "step": 16135
    },
    {
      "epoch": 8.459119496855346,
      "grad_norm": 0.35204610228538513,
      "learning_rate": 3.572891435310717e-05,
      "loss": 0.5793,
      "num_input_tokens_seen": 10554864,
      "step": 16140
    },
    {
      "epoch": 8.46174004192872,
      "grad_norm": 0.3649756610393524,
      "learning_rate": 3.5718585440826527e-05,
      "loss": 0.3742,
      "num_input_tokens_seen": 10558960,
      "step": 16145
    },
    {
      "epoch": 8.464360587002096,
      "grad_norm": 0.41663873195648193,
      "learning_rate": 3.570825428634009e-05,
      "loss": 0.4568,
      "num_input_tokens_seen": 10563600,
      "step": 16150
    },
    {
      "epoch": 8.466981132075471,
      "grad_norm": 0.5875456929206848,
      "learning_rate": 3.569792089180903e-05,
      "loss": 0.4505,
      "num_input_tokens_seen": 10566480,
      "step": 16155
    },
    {
      "epoch": 8.469601677148846,
      "grad_norm": 0.3682287037372589,
      "learning_rate": 3.568758525939497e-05,
      "loss": 0.3571,
      "num_input_tokens_seen": 10569616,
      "step": 16160
    },
    {
      "epoch": 8.472222222222221,
      "grad_norm": 0.44650495052337646,
      "learning_rate": 3.5677247391259995e-05,
      "loss": 0.5272,
      "num_input_tokens_seen": 10572336,
      "step": 16165
    },
    {
      "epoch": 8.474842767295598,
      "grad_norm": 0.5644112825393677,
      "learning_rate": 3.566690728956668e-05,
      "loss": 0.5437,
      "num_input_tokens_seen": 10574576,
      "step": 16170
    },
    {
      "epoch": 8.477463312368974,
      "grad_norm": 0.32973167300224304,
      "learning_rate": 3.5656564956478056e-05,
      "loss": 0.4003,
      "num_input_tokens_seen": 10578640,
      "step": 16175
    },
    {
      "epoch": 8.480083857442349,
      "grad_norm": 0.38780567049980164,
      "learning_rate": 3.5646220394157626e-05,
      "loss": 0.3576,
      "num_input_tokens_seen": 10581392,
      "step": 16180
    },
    {
      "epoch": 8.482704402515724,
      "grad_norm": 0.5871660709381104,
      "learning_rate": 3.563587360476933e-05,
      "loss": 0.6383,
      "num_input_tokens_seen": 10584144,
      "step": 16185
    },
    {
      "epoch": 8.485324947589099,
      "grad_norm": 1.011967420578003,
      "learning_rate": 3.562552459047763e-05,
      "loss": 0.4924,
      "num_input_tokens_seen": 10587600,
      "step": 16190
    },
    {
      "epoch": 8.487945492662474,
      "grad_norm": 0.7027499675750732,
      "learning_rate": 3.561517335344741e-05,
      "loss": 0.4669,
      "num_input_tokens_seen": 10590832,
      "step": 16195
    },
    {
      "epoch": 8.49056603773585,
      "grad_norm": 0.5425897836685181,
      "learning_rate": 3.560481989584402e-05,
      "loss": 0.5025,
      "num_input_tokens_seen": 10593712,
      "step": 16200
    },
    {
      "epoch": 8.493186582809225,
      "grad_norm": 0.3632132112979889,
      "learning_rate": 3.55944642198333e-05,
      "loss": 0.4251,
      "num_input_tokens_seen": 10596336,
      "step": 16205
    },
    {
      "epoch": 8.4958071278826,
      "grad_norm": 0.36318355798721313,
      "learning_rate": 3.5584106327581536e-05,
      "loss": 0.4406,
      "num_input_tokens_seen": 10599536,
      "step": 16210
    },
    {
      "epoch": 8.498427672955975,
      "grad_norm": 0.4738171398639679,
      "learning_rate": 3.557374622125547e-05,
      "loss": 0.2488,
      "num_input_tokens_seen": 10602576,
      "step": 16215
    },
    {
      "epoch": 8.50104821802935,
      "grad_norm": 0.5521293878555298,
      "learning_rate": 3.556338390302234e-05,
      "loss": 0.3884,
      "num_input_tokens_seen": 10605616,
      "step": 16220
    },
    {
      "epoch": 8.503668763102725,
      "grad_norm": 0.6720143556594849,
      "learning_rate": 3.55530193750498e-05,
      "loss": 0.6163,
      "num_input_tokens_seen": 10609040,
      "step": 16225
    },
    {
      "epoch": 8.5062893081761,
      "grad_norm": 0.3725835382938385,
      "learning_rate": 3.5542652639505996e-05,
      "loss": 0.5547,
      "num_input_tokens_seen": 10612304,
      "step": 16230
    },
    {
      "epoch": 8.508909853249476,
      "grad_norm": 0.3361485004425049,
      "learning_rate": 3.5532283698559556e-05,
      "loss": 0.3805,
      "num_input_tokens_seen": 10614928,
      "step": 16235
    },
    {
      "epoch": 8.51153039832285,
      "grad_norm": 0.40327706933021545,
      "learning_rate": 3.55219125543795e-05,
      "loss": 0.576,
      "num_input_tokens_seen": 10618320,
      "step": 16240
    },
    {
      "epoch": 8.514150943396226,
      "grad_norm": 0.6757774949073792,
      "learning_rate": 3.5511539209135396e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 10621296,
      "step": 16245
    },
    {
      "epoch": 8.516771488469601,
      "grad_norm": 0.49610981345176697,
      "learning_rate": 3.550116366499721e-05,
      "loss": 0.4943,
      "num_input_tokens_seen": 10623984,
      "step": 16250
    },
    {
      "epoch": 8.519392033542976,
      "grad_norm": 0.45926329493522644,
      "learning_rate": 3.549078592413538e-05,
      "loss": 0.4433,
      "num_input_tokens_seen": 10628560,
      "step": 16255
    },
    {
      "epoch": 8.522012578616351,
      "grad_norm": 0.5972511172294617,
      "learning_rate": 3.548040598872083e-05,
      "loss": 0.4069,
      "num_input_tokens_seen": 10631856,
      "step": 16260
    },
    {
      "epoch": 8.524633123689728,
      "grad_norm": 0.3014504909515381,
      "learning_rate": 3.5470023860924905e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 10635696,
      "step": 16265
    },
    {
      "epoch": 8.527253668763104,
      "grad_norm": 0.5539090633392334,
      "learning_rate": 3.5459639542919444e-05,
      "loss": 0.4664,
      "num_input_tokens_seen": 10638384,
      "step": 16270
    },
    {
      "epoch": 8.529874213836479,
      "grad_norm": 0.5664703845977783,
      "learning_rate": 3.5449253036876715e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 10641456,
      "step": 16275
    },
    {
      "epoch": 8.532494758909854,
      "grad_norm": 1.2637913227081299,
      "learning_rate": 3.543886434496947e-05,
      "loss": 0.3883,
      "num_input_tokens_seen": 10644336,
      "step": 16280
    },
    {
      "epoch": 8.535115303983229,
      "grad_norm": 0.3379572033882141,
      "learning_rate": 3.542847346937089e-05,
      "loss": 0.3986,
      "num_input_tokens_seen": 10647440,
      "step": 16285
    },
    {
      "epoch": 8.537735849056604,
      "grad_norm": 0.43984362483024597,
      "learning_rate": 3.541808041225464e-05,
      "loss": 0.4259,
      "num_input_tokens_seen": 10649872,
      "step": 16290
    },
    {
      "epoch": 8.54035639412998,
      "grad_norm": 0.38805291056632996,
      "learning_rate": 3.540768517579481e-05,
      "loss": 0.5077,
      "num_input_tokens_seen": 10652976,
      "step": 16295
    },
    {
      "epoch": 8.542976939203355,
      "grad_norm": 0.513370931148529,
      "learning_rate": 3.539728776216599e-05,
      "loss": 0.4202,
      "num_input_tokens_seen": 10656016,
      "step": 16300
    },
    {
      "epoch": 8.54559748427673,
      "grad_norm": 0.5030715465545654,
      "learning_rate": 3.5386888173543185e-05,
      "loss": 0.5064,
      "num_input_tokens_seen": 10659536,
      "step": 16305
    },
    {
      "epoch": 8.548218029350105,
      "grad_norm": 0.43074101209640503,
      "learning_rate": 3.5376486412101875e-05,
      "loss": 0.3599,
      "num_input_tokens_seen": 10662416,
      "step": 16310
    },
    {
      "epoch": 8.55083857442348,
      "grad_norm": 0.27536919713020325,
      "learning_rate": 3.5366082480017994e-05,
      "loss": 0.3864,
      "num_input_tokens_seen": 10666320,
      "step": 16315
    },
    {
      "epoch": 8.553459119496855,
      "grad_norm": 0.5375655889511108,
      "learning_rate": 3.535567637946792e-05,
      "loss": 0.6179,
      "num_input_tokens_seen": 10669136,
      "step": 16320
    },
    {
      "epoch": 8.55607966457023,
      "grad_norm": 0.44050225615501404,
      "learning_rate": 3.534526811262848e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 10672240,
      "step": 16325
    },
    {
      "epoch": 8.558700209643606,
      "grad_norm": 0.6664955615997314,
      "learning_rate": 3.533485768167698e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 10674992,
      "step": 16330
    },
    {
      "epoch": 8.56132075471698,
      "grad_norm": 0.48983263969421387,
      "learning_rate": 3.532444508879117e-05,
      "loss": 0.5006,
      "num_input_tokens_seen": 10677616,
      "step": 16335
    },
    {
      "epoch": 8.563941299790356,
      "grad_norm": 0.42745062708854675,
      "learning_rate": 3.531403033614922e-05,
      "loss": 0.4613,
      "num_input_tokens_seen": 10681136,
      "step": 16340
    },
    {
      "epoch": 8.566561844863731,
      "grad_norm": 0.7338730692863464,
      "learning_rate": 3.530361342592981e-05,
      "loss": 0.6024,
      "num_input_tokens_seen": 10684368,
      "step": 16345
    },
    {
      "epoch": 8.569182389937106,
      "grad_norm": 0.42823532223701477,
      "learning_rate": 3.5293194360312006e-05,
      "loss": 0.512,
      "num_input_tokens_seen": 10687632,
      "step": 16350
    },
    {
      "epoch": 8.571802935010481,
      "grad_norm": 0.24708586931228638,
      "learning_rate": 3.528277314147538e-05,
      "loss": 0.4892,
      "num_input_tokens_seen": 10692336,
      "step": 16355
    },
    {
      "epoch": 8.574423480083858,
      "grad_norm": 0.302383154630661,
      "learning_rate": 3.5272349771599925e-05,
      "loss": 0.4907,
      "num_input_tokens_seen": 10696048,
      "step": 16360
    },
    {
      "epoch": 8.577044025157234,
      "grad_norm": 1.333511471748352,
      "learning_rate": 3.526192425286609e-05,
      "loss": 0.4255,
      "num_input_tokens_seen": 10700304,
      "step": 16365
    },
    {
      "epoch": 8.579664570230609,
      "grad_norm": 0.40275004506111145,
      "learning_rate": 3.525149658745476e-05,
      "loss": 0.4547,
      "num_input_tokens_seen": 10703120,
      "step": 16370
    },
    {
      "epoch": 8.582285115303984,
      "grad_norm": 0.2310740053653717,
      "learning_rate": 3.5241066777547315e-05,
      "loss": 0.4583,
      "num_input_tokens_seen": 10707536,
      "step": 16375
    },
    {
      "epoch": 8.584905660377359,
      "grad_norm": 0.4437675476074219,
      "learning_rate": 3.523063482532554e-05,
      "loss": 0.5264,
      "num_input_tokens_seen": 10710384,
      "step": 16380
    },
    {
      "epoch": 8.587526205450734,
      "grad_norm": 0.5963870882987976,
      "learning_rate": 3.5220200732971665e-05,
      "loss": 0.4471,
      "num_input_tokens_seen": 10713840,
      "step": 16385
    },
    {
      "epoch": 8.59014675052411,
      "grad_norm": 0.44018784165382385,
      "learning_rate": 3.52097645026684e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 10716720,
      "step": 16390
    },
    {
      "epoch": 8.592767295597485,
      "grad_norm": 0.4392853081226349,
      "learning_rate": 3.519932613659889e-05,
      "loss": 0.5019,
      "num_input_tokens_seen": 10720720,
      "step": 16395
    },
    {
      "epoch": 8.59538784067086,
      "grad_norm": 0.5422617197036743,
      "learning_rate": 3.5188885636946696e-05,
      "loss": 0.4381,
      "num_input_tokens_seen": 10723952,
      "step": 16400
    },
    {
      "epoch": 8.598008385744235,
      "grad_norm": 0.5945016145706177,
      "learning_rate": 3.5178443005895884e-05,
      "loss": 0.5533,
      "num_input_tokens_seen": 10726096,
      "step": 16405
    },
    {
      "epoch": 8.60062893081761,
      "grad_norm": 0.30502745509147644,
      "learning_rate": 3.51679982456309e-05,
      "loss": 0.3832,
      "num_input_tokens_seen": 10729456,
      "step": 16410
    },
    {
      "epoch": 8.603249475890985,
      "grad_norm": 0.3649081587791443,
      "learning_rate": 3.51575513583367e-05,
      "loss": 0.3523,
      "num_input_tokens_seen": 10731824,
      "step": 16415
    },
    {
      "epoch": 8.60587002096436,
      "grad_norm": 0.33317211270332336,
      "learning_rate": 3.5147102346198635e-05,
      "loss": 0.3523,
      "num_input_tokens_seen": 10734864,
      "step": 16420
    },
    {
      "epoch": 8.608490566037736,
      "grad_norm": 0.46943724155426025,
      "learning_rate": 3.5136651211402515e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 10737744,
      "step": 16425
    },
    {
      "epoch": 8.61111111111111,
      "grad_norm": 0.5943110585212708,
      "learning_rate": 3.5126197956134614e-05,
      "loss": 0.318,
      "num_input_tokens_seen": 10740656,
      "step": 16430
    },
    {
      "epoch": 8.613731656184486,
      "grad_norm": 0.7686648368835449,
      "learning_rate": 3.5115742582581626e-05,
      "loss": 0.6975,
      "num_input_tokens_seen": 10744688,
      "step": 16435
    },
    {
      "epoch": 8.616352201257861,
      "grad_norm": 0.37346068024635315,
      "learning_rate": 3.51052850929307e-05,
      "loss": 0.5047,
      "num_input_tokens_seen": 10747184,
      "step": 16440
    },
    {
      "epoch": 8.618972746331236,
      "grad_norm": 0.6256058216094971,
      "learning_rate": 3.509482548936941e-05,
      "loss": 0.5869,
      "num_input_tokens_seen": 10750320,
      "step": 16445
    },
    {
      "epoch": 8.621593291404611,
      "grad_norm": 0.4285446107387543,
      "learning_rate": 3.5084363774085804e-05,
      "loss": 0.3817,
      "num_input_tokens_seen": 10753328,
      "step": 16450
    },
    {
      "epoch": 8.624213836477988,
      "grad_norm": 0.47745832800865173,
      "learning_rate": 3.507389994926834e-05,
      "loss": 0.4753,
      "num_input_tokens_seen": 10756208,
      "step": 16455
    },
    {
      "epoch": 8.626834381551364,
      "grad_norm": 0.7638845443725586,
      "learning_rate": 3.5063434017105946e-05,
      "loss": 0.4252,
      "num_input_tokens_seen": 10759888,
      "step": 16460
    },
    {
      "epoch": 8.629454926624739,
      "grad_norm": 0.6288125514984131,
      "learning_rate": 3.505296597978795e-05,
      "loss": 0.4089,
      "num_input_tokens_seen": 10763152,
      "step": 16465
    },
    {
      "epoch": 8.632075471698114,
      "grad_norm": 0.4864775240421295,
      "learning_rate": 3.504249583950417e-05,
      "loss": 0.4623,
      "num_input_tokens_seen": 10767024,
      "step": 16470
    },
    {
      "epoch": 8.634696016771489,
      "grad_norm": 0.39213958382606506,
      "learning_rate": 3.503202359844483e-05,
      "loss": 0.3534,
      "num_input_tokens_seen": 10770096,
      "step": 16475
    },
    {
      "epoch": 8.637316561844864,
      "grad_norm": 0.48799943923950195,
      "learning_rate": 3.50215492588006e-05,
      "loss": 0.52,
      "num_input_tokens_seen": 10774576,
      "step": 16480
    },
    {
      "epoch": 8.63993710691824,
      "grad_norm": 0.3955150544643402,
      "learning_rate": 3.5011072822762605e-05,
      "loss": 0.3493,
      "num_input_tokens_seen": 10777296,
      "step": 16485
    },
    {
      "epoch": 8.642557651991615,
      "grad_norm": 0.2925485670566559,
      "learning_rate": 3.500059429252238e-05,
      "loss": 0.3998,
      "num_input_tokens_seen": 10779696,
      "step": 16490
    },
    {
      "epoch": 8.64517819706499,
      "grad_norm": 0.5601668357849121,
      "learning_rate": 3.499011367027193e-05,
      "loss": 0.3898,
      "num_input_tokens_seen": 10783184,
      "step": 16495
    },
    {
      "epoch": 8.647798742138365,
      "grad_norm": 0.3722945749759674,
      "learning_rate": 3.4979630958203665e-05,
      "loss": 0.6051,
      "num_input_tokens_seen": 10786512,
      "step": 16500
    },
    {
      "epoch": 8.65041928721174,
      "grad_norm": 0.4893682599067688,
      "learning_rate": 3.496914615851046e-05,
      "loss": 0.4146,
      "num_input_tokens_seen": 10789872,
      "step": 16505
    },
    {
      "epoch": 8.653039832285115,
      "grad_norm": 0.34933632612228394,
      "learning_rate": 3.495865927338562e-05,
      "loss": 0.3942,
      "num_input_tokens_seen": 10792944,
      "step": 16510
    },
    {
      "epoch": 8.65566037735849,
      "grad_norm": 0.6470435261726379,
      "learning_rate": 3.494817030502286e-05,
      "loss": 0.4696,
      "num_input_tokens_seen": 10796624,
      "step": 16515
    },
    {
      "epoch": 8.658280922431866,
      "grad_norm": 0.3813227713108063,
      "learning_rate": 3.493767925561638e-05,
      "loss": 0.4408,
      "num_input_tokens_seen": 10800144,
      "step": 16520
    },
    {
      "epoch": 8.66090146750524,
      "grad_norm": 0.6462454199790955,
      "learning_rate": 3.492718612736077e-05,
      "loss": 0.4028,
      "num_input_tokens_seen": 10803152,
      "step": 16525
    },
    {
      "epoch": 8.663522012578616,
      "grad_norm": 0.829195499420166,
      "learning_rate": 3.491669092245108e-05,
      "loss": 0.5066,
      "num_input_tokens_seen": 10805552,
      "step": 16530
    },
    {
      "epoch": 8.666142557651991,
      "grad_norm": 0.52129727602005,
      "learning_rate": 3.4906193643082784e-05,
      "loss": 0.4468,
      "num_input_tokens_seen": 10808880,
      "step": 16535
    },
    {
      "epoch": 8.668763102725366,
      "grad_norm": 0.5018149614334106,
      "learning_rate": 3.4895694291451795e-05,
      "loss": 0.4313,
      "num_input_tokens_seen": 10812240,
      "step": 16540
    },
    {
      "epoch": 8.671383647798741,
      "grad_norm": 1.009770154953003,
      "learning_rate": 3.4885192869754444e-05,
      "loss": 0.4178,
      "num_input_tokens_seen": 10815600,
      "step": 16545
    },
    {
      "epoch": 8.674004192872118,
      "grad_norm": 0.481319785118103,
      "learning_rate": 3.4874689380187537e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 10819120,
      "step": 16550
    },
    {
      "epoch": 8.676624737945493,
      "grad_norm": 0.6257814764976501,
      "learning_rate": 3.486418382494826e-05,
      "loss": 0.3835,
      "num_input_tokens_seen": 10821744,
      "step": 16555
    },
    {
      "epoch": 8.679245283018869,
      "grad_norm": 0.8229389786720276,
      "learning_rate": 3.485367620623426e-05,
      "loss": 0.4209,
      "num_input_tokens_seen": 10824336,
      "step": 16560
    },
    {
      "epoch": 8.681865828092244,
      "grad_norm": 0.6420109272003174,
      "learning_rate": 3.484316652624362e-05,
      "loss": 0.5661,
      "num_input_tokens_seen": 10826672,
      "step": 16565
    },
    {
      "epoch": 8.684486373165619,
      "grad_norm": 0.36516422033309937,
      "learning_rate": 3.483265478717484e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 10830192,
      "step": 16570
    },
    {
      "epoch": 8.687106918238994,
      "grad_norm": 1.2336691617965698,
      "learning_rate": 3.482214099122686e-05,
      "loss": 0.6319,
      "num_input_tokens_seen": 10834096,
      "step": 16575
    },
    {
      "epoch": 8.68972746331237,
      "grad_norm": 0.6235876679420471,
      "learning_rate": 3.481162514059903e-05,
      "loss": 0.4533,
      "num_input_tokens_seen": 10837360,
      "step": 16580
    },
    {
      "epoch": 8.692348008385745,
      "grad_norm": 0.4180731177330017,
      "learning_rate": 3.480110723749116e-05,
      "loss": 0.545,
      "num_input_tokens_seen": 10841168,
      "step": 16585
    },
    {
      "epoch": 8.69496855345912,
      "grad_norm": 0.5820276141166687,
      "learning_rate": 3.479058728410348e-05,
      "loss": 0.3356,
      "num_input_tokens_seen": 10844784,
      "step": 16590
    },
    {
      "epoch": 8.697589098532495,
      "grad_norm": 0.5806328654289246,
      "learning_rate": 3.478006528263663e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 10847376,
      "step": 16595
    },
    {
      "epoch": 8.70020964360587,
      "grad_norm": 0.5581212043762207,
      "learning_rate": 3.4769541235291704e-05,
      "loss": 0.3883,
      "num_input_tokens_seen": 10850000,
      "step": 16600
    },
    {
      "epoch": 8.702830188679245,
      "grad_norm": 0.5136079788208008,
      "learning_rate": 3.475901514427021e-05,
      "loss": 0.3626,
      "num_input_tokens_seen": 10852528,
      "step": 16605
    },
    {
      "epoch": 8.70545073375262,
      "grad_norm": 0.9046577215194702,
      "learning_rate": 3.4748487011774075e-05,
      "loss": 0.4223,
      "num_input_tokens_seen": 10855312,
      "step": 16610
    },
    {
      "epoch": 8.708071278825996,
      "grad_norm": 0.2944584786891937,
      "learning_rate": 3.473795684000569e-05,
      "loss": 0.5972,
      "num_input_tokens_seen": 10859376,
      "step": 16615
    },
    {
      "epoch": 8.71069182389937,
      "grad_norm": 0.478141188621521,
      "learning_rate": 3.4727424631167815e-05,
      "loss": 0.4936,
      "num_input_tokens_seen": 10862224,
      "step": 16620
    },
    {
      "epoch": 8.713312368972746,
      "grad_norm": 0.6861209273338318,
      "learning_rate": 3.471689038746369e-05,
      "loss": 0.5913,
      "num_input_tokens_seen": 10865104,
      "step": 16625
    },
    {
      "epoch": 8.715932914046121,
      "grad_norm": 0.47035831212997437,
      "learning_rate": 3.470635411109695e-05,
      "loss": 0.3898,
      "num_input_tokens_seen": 10867792,
      "step": 16630
    },
    {
      "epoch": 8.718553459119496,
      "grad_norm": 0.6856486797332764,
      "learning_rate": 3.469581580427167e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 10870768,
      "step": 16635
    },
    {
      "epoch": 8.721174004192871,
      "grad_norm": 0.3574945032596588,
      "learning_rate": 3.4685275469192335e-05,
      "loss": 0.3183,
      "num_input_tokens_seen": 10875056,
      "step": 16640
    },
    {
      "epoch": 8.723794549266248,
      "grad_norm": 0.6765676140785217,
      "learning_rate": 3.467473310806387e-05,
      "loss": 0.5065,
      "num_input_tokens_seen": 10878288,
      "step": 16645
    },
    {
      "epoch": 8.726415094339622,
      "grad_norm": 0.5679111480712891,
      "learning_rate": 3.466418872309161e-05,
      "loss": 0.4851,
      "num_input_tokens_seen": 10880720,
      "step": 16650
    },
    {
      "epoch": 8.729035639412999,
      "grad_norm": 0.3960740268230438,
      "learning_rate": 3.465364231648133e-05,
      "loss": 0.281,
      "num_input_tokens_seen": 10884528,
      "step": 16655
    },
    {
      "epoch": 8.731656184486374,
      "grad_norm": 0.5197502374649048,
      "learning_rate": 3.464309389043921e-05,
      "loss": 0.5541,
      "num_input_tokens_seen": 10886960,
      "step": 16660
    },
    {
      "epoch": 8.734276729559749,
      "grad_norm": 0.33141034841537476,
      "learning_rate": 3.4632543447171856e-05,
      "loss": 0.4619,
      "num_input_tokens_seen": 10889904,
      "step": 16665
    },
    {
      "epoch": 8.736897274633124,
      "grad_norm": 0.4844754934310913,
      "learning_rate": 3.4621990988886315e-05,
      "loss": 0.5493,
      "num_input_tokens_seen": 10893072,
      "step": 16670
    },
    {
      "epoch": 8.7395178197065,
      "grad_norm": 0.37747329473495483,
      "learning_rate": 3.461143651779003e-05,
      "loss": 0.4948,
      "num_input_tokens_seen": 10897392,
      "step": 16675
    },
    {
      "epoch": 8.742138364779874,
      "grad_norm": 0.35389888286590576,
      "learning_rate": 3.4600880036090876e-05,
      "loss": 0.5491,
      "num_input_tokens_seen": 10901008,
      "step": 16680
    },
    {
      "epoch": 8.74475890985325,
      "grad_norm": 0.4320453107357025,
      "learning_rate": 3.459032154599715e-05,
      "loss": 0.5186,
      "num_input_tokens_seen": 10904848,
      "step": 16685
    },
    {
      "epoch": 8.747379454926625,
      "grad_norm": 0.6827284097671509,
      "learning_rate": 3.457976104971757e-05,
      "loss": 0.459,
      "num_input_tokens_seen": 10908528,
      "step": 16690
    },
    {
      "epoch": 8.75,
      "grad_norm": 0.45149096846580505,
      "learning_rate": 3.4569198549461276e-05,
      "loss": 0.5634,
      "num_input_tokens_seen": 10912080,
      "step": 16695
    },
    {
      "epoch": 8.752620545073375,
      "grad_norm": 0.3748760223388672,
      "learning_rate": 3.455863404743781e-05,
      "loss": 0.4719,
      "num_input_tokens_seen": 10916528,
      "step": 16700
    },
    {
      "epoch": 8.75524109014675,
      "grad_norm": 0.5180088877677917,
      "learning_rate": 3.454806754585716e-05,
      "loss": 0.4266,
      "num_input_tokens_seen": 10920688,
      "step": 16705
    },
    {
      "epoch": 8.757861635220126,
      "grad_norm": 0.6153659224510193,
      "learning_rate": 3.453749904692971e-05,
      "loss": 0.4397,
      "num_input_tokens_seen": 10923600,
      "step": 16710
    },
    {
      "epoch": 8.7604821802935,
      "grad_norm": 0.3232363760471344,
      "learning_rate": 3.452692855286626e-05,
      "loss": 0.3796,
      "num_input_tokens_seen": 10926544,
      "step": 16715
    },
    {
      "epoch": 8.763102725366876,
      "grad_norm": 0.34669825434684753,
      "learning_rate": 3.4516356065878055e-05,
      "loss": 0.3833,
      "num_input_tokens_seen": 10929392,
      "step": 16720
    },
    {
      "epoch": 8.765723270440251,
      "grad_norm": 0.6900418996810913,
      "learning_rate": 3.450578158817671e-05,
      "loss": 0.4197,
      "num_input_tokens_seen": 10933456,
      "step": 16725
    },
    {
      "epoch": 8.768343815513626,
      "grad_norm": 0.6356337070465088,
      "learning_rate": 3.4495205121974325e-05,
      "loss": 0.594,
      "num_input_tokens_seen": 10936336,
      "step": 16730
    },
    {
      "epoch": 8.770964360587001,
      "grad_norm": 0.44267264008522034,
      "learning_rate": 3.448462666948335e-05,
      "loss": 0.4841,
      "num_input_tokens_seen": 10939696,
      "step": 16735
    },
    {
      "epoch": 8.773584905660378,
      "grad_norm": 0.5591315627098083,
      "learning_rate": 3.4474046232916674e-05,
      "loss": 0.3626,
      "num_input_tokens_seen": 10942192,
      "step": 16740
    },
    {
      "epoch": 8.776205450733752,
      "grad_norm": 0.5101111531257629,
      "learning_rate": 3.4463463814487616e-05,
      "loss": 0.2909,
      "num_input_tokens_seen": 10944624,
      "step": 16745
    },
    {
      "epoch": 8.778825995807129,
      "grad_norm": 0.2621718645095825,
      "learning_rate": 3.44528794164099e-05,
      "loss": 0.3876,
      "num_input_tokens_seen": 10947952,
      "step": 16750
    },
    {
      "epoch": 8.781446540880504,
      "grad_norm": 0.697773277759552,
      "learning_rate": 3.4442293040897646e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 10951440,
      "step": 16755
    },
    {
      "epoch": 8.784067085953879,
      "grad_norm": 0.3686578571796417,
      "learning_rate": 3.443170469016542e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 10954224,
      "step": 16760
    },
    {
      "epoch": 8.786687631027254,
      "grad_norm": 1.0346537828445435,
      "learning_rate": 3.442111436642817e-05,
      "loss": 0.4651,
      "num_input_tokens_seen": 10957392,
      "step": 16765
    },
    {
      "epoch": 8.78930817610063,
      "grad_norm": 0.29206183552742004,
      "learning_rate": 3.441052207190127e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 10961136,
      "step": 16770
    },
    {
      "epoch": 8.791928721174004,
      "grad_norm": 0.3754987418651581,
      "learning_rate": 3.4399927808800526e-05,
      "loss": 0.4841,
      "num_input_tokens_seen": 10964272,
      "step": 16775
    },
    {
      "epoch": 8.79454926624738,
      "grad_norm": 0.3947349190711975,
      "learning_rate": 3.438933157934211e-05,
      "loss": 0.3563,
      "num_input_tokens_seen": 10967344,
      "step": 16780
    },
    {
      "epoch": 8.797169811320755,
      "grad_norm": 0.5438418388366699,
      "learning_rate": 3.4378733385742653e-05,
      "loss": 0.3985,
      "num_input_tokens_seen": 10970800,
      "step": 16785
    },
    {
      "epoch": 8.79979035639413,
      "grad_norm": 0.2698773443698883,
      "learning_rate": 3.436813323021918e-05,
      "loss": 0.4658,
      "num_input_tokens_seen": 10974544,
      "step": 16790
    },
    {
      "epoch": 8.802410901467505,
      "grad_norm": 1.017655611038208,
      "learning_rate": 3.43575311149891e-05,
      "loss": 0.5076,
      "num_input_tokens_seen": 10976784,
      "step": 16795
    },
    {
      "epoch": 8.80503144654088,
      "grad_norm": 0.3646019399166107,
      "learning_rate": 3.434692704227028e-05,
      "loss": 0.5206,
      "num_input_tokens_seen": 10980400,
      "step": 16800
    },
    {
      "epoch": 8.807651991614255,
      "grad_norm": 0.42502716183662415,
      "learning_rate": 3.4336321014280944e-05,
      "loss": 0.3384,
      "num_input_tokens_seen": 10986160,
      "step": 16805
    },
    {
      "epoch": 8.81027253668763,
      "grad_norm": 0.5220676064491272,
      "learning_rate": 3.432571303323977e-05,
      "loss": 0.4825,
      "num_input_tokens_seen": 10989584,
      "step": 16810
    },
    {
      "epoch": 8.812893081761006,
      "grad_norm": 0.7311035394668579,
      "learning_rate": 3.431510310136583e-05,
      "loss": 0.5358,
      "num_input_tokens_seen": 10992464,
      "step": 16815
    },
    {
      "epoch": 8.815513626834381,
      "grad_norm": 0.5793176889419556,
      "learning_rate": 3.4304491220878585e-05,
      "loss": 0.5508,
      "num_input_tokens_seen": 10995472,
      "step": 16820
    },
    {
      "epoch": 8.818134171907756,
      "grad_norm": 0.8189115524291992,
      "learning_rate": 3.429387739399793e-05,
      "loss": 0.4307,
      "num_input_tokens_seen": 10999088,
      "step": 16825
    },
    {
      "epoch": 8.820754716981131,
      "grad_norm": 0.3841702938079834,
      "learning_rate": 3.428326162294416e-05,
      "loss": 0.4639,
      "num_input_tokens_seen": 11002000,
      "step": 16830
    },
    {
      "epoch": 8.823375262054507,
      "grad_norm": 0.7318094968795776,
      "learning_rate": 3.427264390993795e-05,
      "loss": 0.5992,
      "num_input_tokens_seen": 11005360,
      "step": 16835
    },
    {
      "epoch": 8.825995807127882,
      "grad_norm": 0.3715847134590149,
      "learning_rate": 3.426202425720044e-05,
      "loss": 0.3759,
      "num_input_tokens_seen": 11008272,
      "step": 16840
    },
    {
      "epoch": 8.828616352201259,
      "grad_norm": 0.4010026156902313,
      "learning_rate": 3.425140266695312e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 11011760,
      "step": 16845
    },
    {
      "epoch": 8.831236897274634,
      "grad_norm": 0.4772785007953644,
      "learning_rate": 3.4240779141417905e-05,
      "loss": 0.4933,
      "num_input_tokens_seen": 11014736,
      "step": 16850
    },
    {
      "epoch": 8.833857442348009,
      "grad_norm": 0.3337061107158661,
      "learning_rate": 3.423015368281711e-05,
      "loss": 0.3987,
      "num_input_tokens_seen": 11018128,
      "step": 16855
    },
    {
      "epoch": 8.836477987421384,
      "grad_norm": 0.4559265971183777,
      "learning_rate": 3.421952629337348e-05,
      "loss": 0.3985,
      "num_input_tokens_seen": 11022032,
      "step": 16860
    },
    {
      "epoch": 8.83909853249476,
      "grad_norm": 0.5761765241622925,
      "learning_rate": 3.420889697531013e-05,
      "loss": 0.4261,
      "num_input_tokens_seen": 11024944,
      "step": 16865
    },
    {
      "epoch": 8.841719077568134,
      "grad_norm": 0.4402844309806824,
      "learning_rate": 3.419826573085058e-05,
      "loss": 0.416,
      "num_input_tokens_seen": 11028272,
      "step": 16870
    },
    {
      "epoch": 8.84433962264151,
      "grad_norm": 0.9159645438194275,
      "learning_rate": 3.418763256221879e-05,
      "loss": 0.4697,
      "num_input_tokens_seen": 11031088,
      "step": 16875
    },
    {
      "epoch": 8.846960167714885,
      "grad_norm": 0.5386902689933777,
      "learning_rate": 3.4176997471639085e-05,
      "loss": 0.4975,
      "num_input_tokens_seen": 11033456,
      "step": 16880
    },
    {
      "epoch": 8.84958071278826,
      "grad_norm": 0.3187049329280853,
      "learning_rate": 3.4166360461336206e-05,
      "loss": 0.48,
      "num_input_tokens_seen": 11036976,
      "step": 16885
    },
    {
      "epoch": 8.852201257861635,
      "grad_norm": 0.3924543559551239,
      "learning_rate": 3.41557215335353e-05,
      "loss": 0.4576,
      "num_input_tokens_seen": 11039472,
      "step": 16890
    },
    {
      "epoch": 8.85482180293501,
      "grad_norm": 0.4467761218547821,
      "learning_rate": 3.4145080690461887e-05,
      "loss": 0.4283,
      "num_input_tokens_seen": 11042064,
      "step": 16895
    },
    {
      "epoch": 8.857442348008385,
      "grad_norm": 0.38904470205307007,
      "learning_rate": 3.4134437934341936e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 11045232,
      "step": 16900
    },
    {
      "epoch": 8.86006289308176,
      "grad_norm": 0.36226457357406616,
      "learning_rate": 3.4123793267401784e-05,
      "loss": 0.413,
      "num_input_tokens_seen": 11048784,
      "step": 16905
    },
    {
      "epoch": 8.862683438155136,
      "grad_norm": 0.371978759765625,
      "learning_rate": 3.411314669186816e-05,
      "loss": 0.4968,
      "num_input_tokens_seen": 11052528,
      "step": 16910
    },
    {
      "epoch": 8.865303983228511,
      "grad_norm": 0.7619270086288452,
      "learning_rate": 3.410249820996823e-05,
      "loss": 0.6948,
      "num_input_tokens_seen": 11055824,
      "step": 16915
    },
    {
      "epoch": 8.867924528301886,
      "grad_norm": 0.398783415555954,
      "learning_rate": 3.4091847823929513e-05,
      "loss": 0.319,
      "num_input_tokens_seen": 11059504,
      "step": 16920
    },
    {
      "epoch": 8.870545073375261,
      "grad_norm": 0.31141868233680725,
      "learning_rate": 3.4081195535979946e-05,
      "loss": 0.4368,
      "num_input_tokens_seen": 11062544,
      "step": 16925
    },
    {
      "epoch": 8.873165618448636,
      "grad_norm": 0.3199068009853363,
      "learning_rate": 3.40705413483479e-05,
      "loss": 0.4194,
      "num_input_tokens_seen": 11065584,
      "step": 16930
    },
    {
      "epoch": 8.875786163522012,
      "grad_norm": 0.5181244015693665,
      "learning_rate": 3.405988526326207e-05,
      "loss": 0.3735,
      "num_input_tokens_seen": 11068336,
      "step": 16935
    },
    {
      "epoch": 8.878406708595389,
      "grad_norm": 0.6274802684783936,
      "learning_rate": 3.40492272829516e-05,
      "loss": 0.4851,
      "num_input_tokens_seen": 11071184,
      "step": 16940
    },
    {
      "epoch": 8.881027253668764,
      "grad_norm": 0.6836819052696228,
      "learning_rate": 3.403856740964604e-05,
      "loss": 0.2823,
      "num_input_tokens_seen": 11074320,
      "step": 16945
    },
    {
      "epoch": 8.883647798742139,
      "grad_norm": 0.38782942295074463,
      "learning_rate": 3.4027905645575275e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 11077168,
      "step": 16950
    },
    {
      "epoch": 8.886268343815514,
      "grad_norm": 0.7004958987236023,
      "learning_rate": 3.401724199296966e-05,
      "loss": 0.4474,
      "num_input_tokens_seen": 11080080,
      "step": 16955
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 0.4124552309513092,
      "learning_rate": 3.400657645405989e-05,
      "loss": 0.3558,
      "num_input_tokens_seen": 11083408,
      "step": 16960
    },
    {
      "epoch": 8.891509433962264,
      "grad_norm": 0.4962787926197052,
      "learning_rate": 3.3995909031077065e-05,
      "loss": 0.5589,
      "num_input_tokens_seen": 11087184,
      "step": 16965
    },
    {
      "epoch": 8.89412997903564,
      "grad_norm": 0.39066585898399353,
      "learning_rate": 3.398523972625272e-05,
      "loss": 0.3832,
      "num_input_tokens_seen": 11090864,
      "step": 16970
    },
    {
      "epoch": 8.896750524109015,
      "grad_norm": 0.3779095709323883,
      "learning_rate": 3.3974568541818723e-05,
      "loss": 0.3999,
      "num_input_tokens_seen": 11094288,
      "step": 16975
    },
    {
      "epoch": 8.89937106918239,
      "grad_norm": 0.33611100912094116,
      "learning_rate": 3.396389548000738e-05,
      "loss": 0.2966,
      "num_input_tokens_seen": 11096656,
      "step": 16980
    },
    {
      "epoch": 8.901991614255765,
      "grad_norm": 0.26834356784820557,
      "learning_rate": 3.395322054305137e-05,
      "loss": 0.4253,
      "num_input_tokens_seen": 11100144,
      "step": 16985
    },
    {
      "epoch": 8.90461215932914,
      "grad_norm": 0.3101027309894562,
      "learning_rate": 3.394254373318376e-05,
      "loss": 0.397,
      "num_input_tokens_seen": 11103440,
      "step": 16990
    },
    {
      "epoch": 8.907232704402515,
      "grad_norm": 0.39653536677360535,
      "learning_rate": 3.3931865052638026e-05,
      "loss": 0.4144,
      "num_input_tokens_seen": 11107056,
      "step": 16995
    },
    {
      "epoch": 8.90985324947589,
      "grad_norm": 0.5857251286506653,
      "learning_rate": 3.392118450364802e-05,
      "loss": 0.4204,
      "num_input_tokens_seen": 11110640,
      "step": 17000
    },
    {
      "epoch": 8.912473794549266,
      "grad_norm": 0.08797422796487808,
      "learning_rate": 3.391050208844799e-05,
      "loss": 0.3205,
      "num_input_tokens_seen": 11117008,
      "step": 17005
    },
    {
      "epoch": 8.915094339622641,
      "grad_norm": 0.43637046217918396,
      "learning_rate": 3.389981780927259e-05,
      "loss": 0.3957,
      "num_input_tokens_seen": 11120176,
      "step": 17010
    },
    {
      "epoch": 8.917714884696016,
      "grad_norm": 0.5221354961395264,
      "learning_rate": 3.3889131668356834e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 11123504,
      "step": 17015
    },
    {
      "epoch": 8.920335429769391,
      "grad_norm": 0.5891143083572388,
      "learning_rate": 3.387844366793614e-05,
      "loss": 0.4935,
      "num_input_tokens_seen": 11126000,
      "step": 17020
    },
    {
      "epoch": 8.922955974842766,
      "grad_norm": 0.32000282406806946,
      "learning_rate": 3.3867753810246326e-05,
      "loss": 0.4157,
      "num_input_tokens_seen": 11129584,
      "step": 17025
    },
    {
      "epoch": 8.925576519916142,
      "grad_norm": 0.4571104943752289,
      "learning_rate": 3.3857062097523586e-05,
      "loss": 0.4741,
      "num_input_tokens_seen": 11132240,
      "step": 17030
    },
    {
      "epoch": 8.928197064989519,
      "grad_norm": 0.4943161904811859,
      "learning_rate": 3.38463685320045e-05,
      "loss": 0.483,
      "num_input_tokens_seen": 11135120,
      "step": 17035
    },
    {
      "epoch": 8.930817610062894,
      "grad_norm": 0.9015825986862183,
      "learning_rate": 3.3835673115926045e-05,
      "loss": 0.3297,
      "num_input_tokens_seen": 11137968,
      "step": 17040
    },
    {
      "epoch": 8.933438155136269,
      "grad_norm": 0.34895774722099304,
      "learning_rate": 3.382497585152557e-05,
      "loss": 0.3595,
      "num_input_tokens_seen": 11140368,
      "step": 17045
    },
    {
      "epoch": 8.936058700209644,
      "grad_norm": 0.42501118779182434,
      "learning_rate": 3.3814276741040835e-05,
      "loss": 0.5283,
      "num_input_tokens_seen": 11145168,
      "step": 17050
    },
    {
      "epoch": 8.93867924528302,
      "grad_norm": 0.4409411549568176,
      "learning_rate": 3.380357578670996e-05,
      "loss": 0.4841,
      "num_input_tokens_seen": 11149008,
      "step": 17055
    },
    {
      "epoch": 8.941299790356394,
      "grad_norm": 0.48845505714416504,
      "learning_rate": 3.379287299077146e-05,
      "loss": 0.4235,
      "num_input_tokens_seen": 11156336,
      "step": 17060
    },
    {
      "epoch": 8.94392033542977,
      "grad_norm": 0.5463280081748962,
      "learning_rate": 3.3782168355464263e-05,
      "loss": 0.5448,
      "num_input_tokens_seen": 11159152,
      "step": 17065
    },
    {
      "epoch": 8.946540880503145,
      "grad_norm": 0.41188767552375793,
      "learning_rate": 3.377146188302762e-05,
      "loss": 0.3969,
      "num_input_tokens_seen": 11162096,
      "step": 17070
    },
    {
      "epoch": 8.94916142557652,
      "grad_norm": 0.5055385231971741,
      "learning_rate": 3.376075357570124e-05,
      "loss": 0.4452,
      "num_input_tokens_seen": 11165200,
      "step": 17075
    },
    {
      "epoch": 8.951781970649895,
      "grad_norm": 0.7374429106712341,
      "learning_rate": 3.3750043435725144e-05,
      "loss": 0.5673,
      "num_input_tokens_seen": 11168368,
      "step": 17080
    },
    {
      "epoch": 8.95440251572327,
      "grad_norm": 0.3147520124912262,
      "learning_rate": 3.373933146533979e-05,
      "loss": 0.52,
      "num_input_tokens_seen": 11172048,
      "step": 17085
    },
    {
      "epoch": 8.957023060796645,
      "grad_norm": 0.32567980885505676,
      "learning_rate": 3.3728617666786e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 11175600,
      "step": 17090
    },
    {
      "epoch": 8.95964360587002,
      "grad_norm": 0.8368294835090637,
      "learning_rate": 3.371790204230497e-05,
      "loss": 0.4161,
      "num_input_tokens_seen": 11178576,
      "step": 17095
    },
    {
      "epoch": 8.962264150943396,
      "grad_norm": 0.44317808747291565,
      "learning_rate": 3.3707184594138295e-05,
      "loss": 0.5161,
      "num_input_tokens_seen": 11181168,
      "step": 17100
    },
    {
      "epoch": 8.964884696016771,
      "grad_norm": 0.6742874979972839,
      "learning_rate": 3.369646532452795e-05,
      "loss": 0.3978,
      "num_input_tokens_seen": 11184400,
      "step": 17105
    },
    {
      "epoch": 8.967505241090146,
      "grad_norm": 0.45073050260543823,
      "learning_rate": 3.3685744235716254e-05,
      "loss": 0.493,
      "num_input_tokens_seen": 11187536,
      "step": 17110
    },
    {
      "epoch": 8.970125786163521,
      "grad_norm": 0.4976419508457184,
      "learning_rate": 3.367502132994598e-05,
      "loss": 0.4213,
      "num_input_tokens_seen": 11191568,
      "step": 17115
    },
    {
      "epoch": 8.972746331236896,
      "grad_norm": 1.1604808568954468,
      "learning_rate": 3.366429660946019e-05,
      "loss": 0.3781,
      "num_input_tokens_seen": 11195024,
      "step": 17120
    },
    {
      "epoch": 8.975366876310272,
      "grad_norm": 0.6521639823913574,
      "learning_rate": 3.365357007650241e-05,
      "loss": 0.4926,
      "num_input_tokens_seen": 11198480,
      "step": 17125
    },
    {
      "epoch": 8.977987421383649,
      "grad_norm": 0.24800848960876465,
      "learning_rate": 3.364284173331649e-05,
      "loss": 0.4819,
      "num_input_tokens_seen": 11201712,
      "step": 17130
    },
    {
      "epoch": 8.980607966457024,
      "grad_norm": 0.3001781404018402,
      "learning_rate": 3.3632111582146686e-05,
      "loss": 0.3903,
      "num_input_tokens_seen": 11204720,
      "step": 17135
    },
    {
      "epoch": 8.983228511530399,
      "grad_norm": 0.6315777897834778,
      "learning_rate": 3.362137962523761e-05,
      "loss": 0.4423,
      "num_input_tokens_seen": 11207440,
      "step": 17140
    },
    {
      "epoch": 8.985849056603774,
      "grad_norm": 0.41439682245254517,
      "learning_rate": 3.361064586483429e-05,
      "loss": 0.4346,
      "num_input_tokens_seen": 11210384,
      "step": 17145
    },
    {
      "epoch": 8.98846960167715,
      "grad_norm": 0.5424720644950867,
      "learning_rate": 3.359991030318208e-05,
      "loss": 0.441,
      "num_input_tokens_seen": 11213392,
      "step": 17150
    },
    {
      "epoch": 8.991090146750524,
      "grad_norm": 0.5799633264541626,
      "learning_rate": 3.358917294252675e-05,
      "loss": 0.52,
      "num_input_tokens_seen": 11216016,
      "step": 17155
    },
    {
      "epoch": 8.9937106918239,
      "grad_norm": 0.4158553183078766,
      "learning_rate": 3.357843378511443e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 11218160,
      "step": 17160
    },
    {
      "epoch": 8.996331236897275,
      "grad_norm": 0.448011189699173,
      "learning_rate": 3.3567692833191635e-05,
      "loss": 0.6856,
      "num_input_tokens_seen": 11222000,
      "step": 17165
    },
    {
      "epoch": 8.99895178197065,
      "grad_norm": 0.5693494081497192,
      "learning_rate": 3.355695008900525e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 11224816,
      "step": 17170
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.4779369831085205,
      "eval_runtime": 9.721,
      "eval_samples_per_second": 87.234,
      "eval_steps_per_second": 21.808,
      "num_input_tokens_seen": 11225416,
      "step": 17172
    },
    {
      "epoch": 9.001572327044025,
      "grad_norm": 0.6113901734352112,
      "learning_rate": 3.354620555480251e-05,
      "loss": 0.569,
      "num_input_tokens_seen": 11227112,
      "step": 17175
    },
    {
      "epoch": 9.0041928721174,
      "grad_norm": 0.4026243984699249,
      "learning_rate": 3.353545923283108e-05,
      "loss": 0.405,
      "num_input_tokens_seen": 11229544,
      "step": 17180
    },
    {
      "epoch": 9.006813417190775,
      "grad_norm": 0.37351930141448975,
      "learning_rate": 3.352471112533895e-05,
      "loss": 0.4769,
      "num_input_tokens_seen": 11233640,
      "step": 17185
    },
    {
      "epoch": 9.00943396226415,
      "grad_norm": 0.4686587154865265,
      "learning_rate": 3.351396123457451e-05,
      "loss": 0.3696,
      "num_input_tokens_seen": 11236552,
      "step": 17190
    },
    {
      "epoch": 9.012054507337526,
      "grad_norm": 0.5990042090415955,
      "learning_rate": 3.350320956278652e-05,
      "loss": 0.4093,
      "num_input_tokens_seen": 11240008,
      "step": 17195
    },
    {
      "epoch": 9.014675052410901,
      "grad_norm": 0.5490583777427673,
      "learning_rate": 3.349245611222409e-05,
      "loss": 0.4542,
      "num_input_tokens_seen": 11243176,
      "step": 17200
    },
    {
      "epoch": 9.017295597484276,
      "grad_norm": 0.4895869195461273,
      "learning_rate": 3.348170088513673e-05,
      "loss": 0.4333,
      "num_input_tokens_seen": 11246568,
      "step": 17205
    },
    {
      "epoch": 9.019916142557651,
      "grad_norm": 0.25065878033638,
      "learning_rate": 3.347094388377431e-05,
      "loss": 0.4187,
      "num_input_tokens_seen": 11251080,
      "step": 17210
    },
    {
      "epoch": 9.022536687631026,
      "grad_norm": 0.6463639736175537,
      "learning_rate": 3.3460185110387076e-05,
      "loss": 0.6393,
      "num_input_tokens_seen": 11253256,
      "step": 17215
    },
    {
      "epoch": 9.025157232704403,
      "grad_norm": 0.3770753741264343,
      "learning_rate": 3.344942456722563e-05,
      "loss": 0.4415,
      "num_input_tokens_seen": 11256936,
      "step": 17220
    },
    {
      "epoch": 9.027777777777779,
      "grad_norm": 0.4202304184436798,
      "learning_rate": 3.343866225654096e-05,
      "loss": 0.4653,
      "num_input_tokens_seen": 11259784,
      "step": 17225
    },
    {
      "epoch": 9.030398322851154,
      "grad_norm": 0.36649179458618164,
      "learning_rate": 3.342789818058443e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 11262728,
      "step": 17230
    },
    {
      "epoch": 9.033018867924529,
      "grad_norm": 0.43599414825439453,
      "learning_rate": 3.341713234160774e-05,
      "loss": 0.4168,
      "num_input_tokens_seen": 11265640,
      "step": 17235
    },
    {
      "epoch": 9.035639412997904,
      "grad_norm": 0.30640462040901184,
      "learning_rate": 3.3406364741862986e-05,
      "loss": 0.3905,
      "num_input_tokens_seen": 11270344,
      "step": 17240
    },
    {
      "epoch": 9.03825995807128,
      "grad_norm": 1.7170212268829346,
      "learning_rate": 3.3395595383602644e-05,
      "loss": 0.5707,
      "num_input_tokens_seen": 11272584,
      "step": 17245
    },
    {
      "epoch": 9.040880503144654,
      "grad_norm": 0.4314677119255066,
      "learning_rate": 3.338482426907953e-05,
      "loss": 0.4801,
      "num_input_tokens_seen": 11276872,
      "step": 17250
    },
    {
      "epoch": 9.04350104821803,
      "grad_norm": 0.38165178894996643,
      "learning_rate": 3.337405140054683e-05,
      "loss": 0.4517,
      "num_input_tokens_seen": 11279688,
      "step": 17255
    },
    {
      "epoch": 9.046121593291405,
      "grad_norm": 0.46501681208610535,
      "learning_rate": 3.336327678025813e-05,
      "loss": 0.4016,
      "num_input_tokens_seen": 11282888,
      "step": 17260
    },
    {
      "epoch": 9.04874213836478,
      "grad_norm": 0.4024915099143982,
      "learning_rate": 3.3352500410467324e-05,
      "loss": 0.5343,
      "num_input_tokens_seen": 11285928,
      "step": 17265
    },
    {
      "epoch": 9.051362683438155,
      "grad_norm": 0.585656464099884,
      "learning_rate": 3.3341722293428726e-05,
      "loss": 0.4192,
      "num_input_tokens_seen": 11289288,
      "step": 17270
    },
    {
      "epoch": 9.05398322851153,
      "grad_norm": 0.3083685040473938,
      "learning_rate": 3.3330942431397e-05,
      "loss": 0.5099,
      "num_input_tokens_seen": 11295976,
      "step": 17275
    },
    {
      "epoch": 9.056603773584905,
      "grad_norm": 0.36109060049057007,
      "learning_rate": 3.332016082662715e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 11299816,
      "step": 17280
    },
    {
      "epoch": 9.05922431865828,
      "grad_norm": 0.23787745833396912,
      "learning_rate": 3.3309377481374584e-05,
      "loss": 0.5298,
      "num_input_tokens_seen": 11303176,
      "step": 17285
    },
    {
      "epoch": 9.061844863731656,
      "grad_norm": 0.5072780251502991,
      "learning_rate": 3.3298592397895046e-05,
      "loss": 0.3393,
      "num_input_tokens_seen": 11306376,
      "step": 17290
    },
    {
      "epoch": 9.064465408805031,
      "grad_norm": 0.37946391105651855,
      "learning_rate": 3.3287805578444645e-05,
      "loss": 0.3453,
      "num_input_tokens_seen": 11310696,
      "step": 17295
    },
    {
      "epoch": 9.067085953878406,
      "grad_norm": 0.8942651748657227,
      "learning_rate": 3.327701702527989e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 11314888,
      "step": 17300
    },
    {
      "epoch": 9.069706498951781,
      "grad_norm": 0.6123968362808228,
      "learning_rate": 3.3266226740657584e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 11317576,
      "step": 17305
    },
    {
      "epoch": 9.072327044025156,
      "grad_norm": 0.4161354601383209,
      "learning_rate": 3.3255434726834954e-05,
      "loss": 0.4638,
      "num_input_tokens_seen": 11320840,
      "step": 17310
    },
    {
      "epoch": 9.074947589098532,
      "grad_norm": 0.4520539939403534,
      "learning_rate": 3.324464098606957e-05,
      "loss": 0.2978,
      "num_input_tokens_seen": 11324040,
      "step": 17315
    },
    {
      "epoch": 9.077568134171909,
      "grad_norm": 0.5523752570152283,
      "learning_rate": 3.323384552061934e-05,
      "loss": 0.5271,
      "num_input_tokens_seen": 11327784,
      "step": 17320
    },
    {
      "epoch": 9.080188679245284,
      "grad_norm": 0.27037718892097473,
      "learning_rate": 3.322304833274257e-05,
      "loss": 0.3991,
      "num_input_tokens_seen": 11331304,
      "step": 17325
    },
    {
      "epoch": 9.082809224318659,
      "grad_norm": 0.4897981882095337,
      "learning_rate": 3.32122494246979e-05,
      "loss": 0.541,
      "num_input_tokens_seen": 11334312,
      "step": 17330
    },
    {
      "epoch": 9.085429769392034,
      "grad_norm": 0.26058655977249146,
      "learning_rate": 3.320144879874434e-05,
      "loss": 0.473,
      "num_input_tokens_seen": 11339144,
      "step": 17335
    },
    {
      "epoch": 9.08805031446541,
      "grad_norm": 0.5150274038314819,
      "learning_rate": 3.319064645714126e-05,
      "loss": 0.3152,
      "num_input_tokens_seen": 11342664,
      "step": 17340
    },
    {
      "epoch": 9.090670859538784,
      "grad_norm": 0.44008538126945496,
      "learning_rate": 3.3179842402148386e-05,
      "loss": 0.3906,
      "num_input_tokens_seen": 11345768,
      "step": 17345
    },
    {
      "epoch": 9.09329140461216,
      "grad_norm": 0.4233700633049011,
      "learning_rate": 3.316903663602581e-05,
      "loss": 0.5839,
      "num_input_tokens_seen": 11349448,
      "step": 17350
    },
    {
      "epoch": 9.095911949685535,
      "grad_norm": 0.4785233736038208,
      "learning_rate": 3.315822916103395e-05,
      "loss": 0.3812,
      "num_input_tokens_seen": 11352744,
      "step": 17355
    },
    {
      "epoch": 9.09853249475891,
      "grad_norm": 1.4763131141662598,
      "learning_rate": 3.3147419979433635e-05,
      "loss": 0.4973,
      "num_input_tokens_seen": 11354920,
      "step": 17360
    },
    {
      "epoch": 9.101153039832285,
      "grad_norm": 0.5889632105827332,
      "learning_rate": 3.313660909348601e-05,
      "loss": 0.3814,
      "num_input_tokens_seen": 11359240,
      "step": 17365
    },
    {
      "epoch": 9.10377358490566,
      "grad_norm": 0.4044691026210785,
      "learning_rate": 3.312579650545259e-05,
      "loss": 0.5056,
      "num_input_tokens_seen": 11362056,
      "step": 17370
    },
    {
      "epoch": 9.106394129979035,
      "grad_norm": 0.5160926580429077,
      "learning_rate": 3.311498221759524e-05,
      "loss": 0.3523,
      "num_input_tokens_seen": 11364648,
      "step": 17375
    },
    {
      "epoch": 9.10901467505241,
      "grad_norm": 0.4002145230770111,
      "learning_rate": 3.310416623217619e-05,
      "loss": 0.3918,
      "num_input_tokens_seen": 11368168,
      "step": 17380
    },
    {
      "epoch": 9.111635220125786,
      "grad_norm": 0.4570760726928711,
      "learning_rate": 3.309334855145803e-05,
      "loss": 0.4748,
      "num_input_tokens_seen": 11370856,
      "step": 17385
    },
    {
      "epoch": 9.114255765199161,
      "grad_norm": 0.5331980586051941,
      "learning_rate": 3.3082529177703684e-05,
      "loss": 0.4519,
      "num_input_tokens_seen": 11373768,
      "step": 17390
    },
    {
      "epoch": 9.116876310272536,
      "grad_norm": 0.5625673532485962,
      "learning_rate": 3.3071708113176435e-05,
      "loss": 0.4943,
      "num_input_tokens_seen": 11377192,
      "step": 17395
    },
    {
      "epoch": 9.119496855345911,
      "grad_norm": 0.4103235602378845,
      "learning_rate": 3.306088536013995e-05,
      "loss": 0.4216,
      "num_input_tokens_seen": 11379912,
      "step": 17400
    },
    {
      "epoch": 9.122117400419286,
      "grad_norm": 0.3962879776954651,
      "learning_rate": 3.3050060920858206e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 11383496,
      "step": 17405
    },
    {
      "epoch": 9.124737945492662,
      "grad_norm": 0.5581718683242798,
      "learning_rate": 3.3039234797595544e-05,
      "loss": 0.587,
      "num_input_tokens_seen": 11386504,
      "step": 17410
    },
    {
      "epoch": 9.127358490566039,
      "grad_norm": 2.5473904609680176,
      "learning_rate": 3.3028406992616695e-05,
      "loss": 0.3752,
      "num_input_tokens_seen": 11388776,
      "step": 17415
    },
    {
      "epoch": 9.129979035639414,
      "grad_norm": 0.43122729659080505,
      "learning_rate": 3.301757750818669e-05,
      "loss": 0.5771,
      "num_input_tokens_seen": 11391976,
      "step": 17420
    },
    {
      "epoch": 9.132599580712789,
      "grad_norm": 0.2859971821308136,
      "learning_rate": 3.300674634657094e-05,
      "loss": 0.4202,
      "num_input_tokens_seen": 11394856,
      "step": 17425
    },
    {
      "epoch": 9.135220125786164,
      "grad_norm": 0.7070423364639282,
      "learning_rate": 3.2995913510035206e-05,
      "loss": 0.4946,
      "num_input_tokens_seen": 11397896,
      "step": 17430
    },
    {
      "epoch": 9.13784067085954,
      "grad_norm": 0.7703961730003357,
      "learning_rate": 3.298507900084558e-05,
      "loss": 0.5207,
      "num_input_tokens_seen": 11401096,
      "step": 17435
    },
    {
      "epoch": 9.140461215932914,
      "grad_norm": 0.7955980896949768,
      "learning_rate": 3.2974242821268524e-05,
      "loss": 0.4217,
      "num_input_tokens_seen": 11404616,
      "step": 17440
    },
    {
      "epoch": 9.14308176100629,
      "grad_norm": 0.4183536469936371,
      "learning_rate": 3.2963404973570855e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 11407848,
      "step": 17445
    },
    {
      "epoch": 9.145702306079665,
      "grad_norm": 0.4075901210308075,
      "learning_rate": 3.2952565460019715e-05,
      "loss": 0.4213,
      "num_input_tokens_seen": 11410760,
      "step": 17450
    },
    {
      "epoch": 9.14832285115304,
      "grad_norm": 0.5657860636711121,
      "learning_rate": 3.29417242828826e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 11413864,
      "step": 17455
    },
    {
      "epoch": 9.150943396226415,
      "grad_norm": 0.373595654964447,
      "learning_rate": 3.2930881444427374e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 11416808,
      "step": 17460
    },
    {
      "epoch": 9.15356394129979,
      "grad_norm": 0.43471717834472656,
      "learning_rate": 3.292003694692223e-05,
      "loss": 0.5274,
      "num_input_tokens_seen": 11419944,
      "step": 17465
    },
    {
      "epoch": 9.156184486373165,
      "grad_norm": 0.4896840453147888,
      "learning_rate": 3.290919079263572e-05,
      "loss": 0.4005,
      "num_input_tokens_seen": 11422728,
      "step": 17470
    },
    {
      "epoch": 9.15880503144654,
      "grad_norm": 0.30569538474082947,
      "learning_rate": 3.2898342983836715e-05,
      "loss": 0.4788,
      "num_input_tokens_seen": 11430984,
      "step": 17475
    },
    {
      "epoch": 9.161425576519916,
      "grad_norm": 0.5823033452033997,
      "learning_rate": 3.288749352279447e-05,
      "loss": 0.4357,
      "num_input_tokens_seen": 11433672,
      "step": 17480
    },
    {
      "epoch": 9.164046121593291,
      "grad_norm": 0.5252336859703064,
      "learning_rate": 3.2876642411778573e-05,
      "loss": 0.4767,
      "num_input_tokens_seen": 11436168,
      "step": 17485
    },
    {
      "epoch": 9.166666666666666,
      "grad_norm": 0.28693512082099915,
      "learning_rate": 3.286578965305894e-05,
      "loss": 0.3285,
      "num_input_tokens_seen": 11439528,
      "step": 17490
    },
    {
      "epoch": 9.169287211740041,
      "grad_norm": 0.6538878679275513,
      "learning_rate": 3.285493524890585e-05,
      "loss": 0.499,
      "num_input_tokens_seen": 11442312,
      "step": 17495
    },
    {
      "epoch": 9.171907756813416,
      "grad_norm": 0.6298129558563232,
      "learning_rate": 3.284407920158992e-05,
      "loss": 0.3961,
      "num_input_tokens_seen": 11445416,
      "step": 17500
    },
    {
      "epoch": 9.174528301886792,
      "grad_norm": 0.6379303336143494,
      "learning_rate": 3.283322151338211e-05,
      "loss": 0.5996,
      "num_input_tokens_seen": 11448712,
      "step": 17505
    },
    {
      "epoch": 9.177148846960169,
      "grad_norm": 0.3836776912212372,
      "learning_rate": 3.282236218655373e-05,
      "loss": 0.4839,
      "num_input_tokens_seen": 11452104,
      "step": 17510
    },
    {
      "epoch": 9.179769392033544,
      "grad_norm": 0.37927600741386414,
      "learning_rate": 3.281150122337642e-05,
      "loss": 0.3172,
      "num_input_tokens_seen": 11456136,
      "step": 17515
    },
    {
      "epoch": 9.182389937106919,
      "grad_norm": 0.4484941065311432,
      "learning_rate": 3.280063862612216e-05,
      "loss": 0.3754,
      "num_input_tokens_seen": 11459272,
      "step": 17520
    },
    {
      "epoch": 9.185010482180294,
      "grad_norm": 0.4128970205783844,
      "learning_rate": 3.2789774397063314e-05,
      "loss": 0.4199,
      "num_input_tokens_seen": 11463208,
      "step": 17525
    },
    {
      "epoch": 9.18763102725367,
      "grad_norm": 0.43379390239715576,
      "learning_rate": 3.277890853847253e-05,
      "loss": 0.4192,
      "num_input_tokens_seen": 11466472,
      "step": 17530
    },
    {
      "epoch": 9.190251572327044,
      "grad_norm": 0.7833365797996521,
      "learning_rate": 3.2768041052622824e-05,
      "loss": 0.5392,
      "num_input_tokens_seen": 11469192,
      "step": 17535
    },
    {
      "epoch": 9.19287211740042,
      "grad_norm": 0.5142669677734375,
      "learning_rate": 3.275717194178755e-05,
      "loss": 0.4121,
      "num_input_tokens_seen": 11472232,
      "step": 17540
    },
    {
      "epoch": 9.195492662473795,
      "grad_norm": 0.41252246499061584,
      "learning_rate": 3.27463012082404e-05,
      "loss": 0.4882,
      "num_input_tokens_seen": 11476040,
      "step": 17545
    },
    {
      "epoch": 9.19811320754717,
      "grad_norm": 0.4745689928531647,
      "learning_rate": 3.2735428854255416e-05,
      "loss": 0.431,
      "num_input_tokens_seen": 11478728,
      "step": 17550
    },
    {
      "epoch": 9.200733752620545,
      "grad_norm": 0.5075179934501648,
      "learning_rate": 3.272455488210697e-05,
      "loss": 0.5197,
      "num_input_tokens_seen": 11481448,
      "step": 17555
    },
    {
      "epoch": 9.20335429769392,
      "grad_norm": 0.3962763845920563,
      "learning_rate": 3.271367929406977e-05,
      "loss": 0.4278,
      "num_input_tokens_seen": 11484424,
      "step": 17560
    },
    {
      "epoch": 9.205974842767295,
      "grad_norm": 0.5628193020820618,
      "learning_rate": 3.270280209241886e-05,
      "loss": 0.4242,
      "num_input_tokens_seen": 11487560,
      "step": 17565
    },
    {
      "epoch": 9.20859538784067,
      "grad_norm": 0.9961385130882263,
      "learning_rate": 3.2691923279429624e-05,
      "loss": 0.4393,
      "num_input_tokens_seen": 11490632,
      "step": 17570
    },
    {
      "epoch": 9.211215932914046,
      "grad_norm": 0.8626353740692139,
      "learning_rate": 3.2681042857377804e-05,
      "loss": 0.4144,
      "num_input_tokens_seen": 11493032,
      "step": 17575
    },
    {
      "epoch": 9.213836477987421,
      "grad_norm": 0.4157293140888214,
      "learning_rate": 3.2670160828539435e-05,
      "loss": 0.3774,
      "num_input_tokens_seen": 11496200,
      "step": 17580
    },
    {
      "epoch": 9.216457023060796,
      "grad_norm": 0.6148793697357178,
      "learning_rate": 3.2659277195190934e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 11498696,
      "step": 17585
    },
    {
      "epoch": 9.219077568134171,
      "grad_norm": 0.5135149359703064,
      "learning_rate": 3.2648391959609026e-05,
      "loss": 0.3879,
      "num_input_tokens_seen": 11502312,
      "step": 17590
    },
    {
      "epoch": 9.221698113207546,
      "grad_norm": 0.2586107850074768,
      "learning_rate": 3.263750512407078e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 11505160,
      "step": 17595
    },
    {
      "epoch": 9.224318658280922,
      "grad_norm": 0.3813594579696655,
      "learning_rate": 3.262661669085359e-05,
      "loss": 0.3526,
      "num_input_tokens_seen": 11508456,
      "step": 17600
    },
    {
      "epoch": 9.226939203354299,
      "grad_norm": 0.49815988540649414,
      "learning_rate": 3.2615726662235205e-05,
      "loss": 0.5057,
      "num_input_tokens_seen": 11511816,
      "step": 17605
    },
    {
      "epoch": 9.229559748427674,
      "grad_norm": 0.4931301474571228,
      "learning_rate": 3.2604835040493676e-05,
      "loss": 0.5073,
      "num_input_tokens_seen": 11514088,
      "step": 17610
    },
    {
      "epoch": 9.232180293501049,
      "grad_norm": 0.4699043929576874,
      "learning_rate": 3.259394182790742e-05,
      "loss": 0.3763,
      "num_input_tokens_seen": 11517096,
      "step": 17615
    },
    {
      "epoch": 9.234800838574424,
      "grad_norm": 0.47724097967147827,
      "learning_rate": 3.2583047026755174e-05,
      "loss": 0.4068,
      "num_input_tokens_seen": 11520168,
      "step": 17620
    },
    {
      "epoch": 9.2374213836478,
      "grad_norm": 0.4104841947555542,
      "learning_rate": 3.2572150639316e-05,
      "loss": 0.5591,
      "num_input_tokens_seen": 11523944,
      "step": 17625
    },
    {
      "epoch": 9.240041928721174,
      "grad_norm": 0.2252703458070755,
      "learning_rate": 3.256125266786931e-05,
      "loss": 0.5886,
      "num_input_tokens_seen": 11527720,
      "step": 17630
    },
    {
      "epoch": 9.24266247379455,
      "grad_norm": 0.3625166118144989,
      "learning_rate": 3.255035311469481e-05,
      "loss": 0.5095,
      "num_input_tokens_seen": 11531304,
      "step": 17635
    },
    {
      "epoch": 9.245283018867925,
      "grad_norm": 0.41845348477363586,
      "learning_rate": 3.2539451982072586e-05,
      "loss": 0.4041,
      "num_input_tokens_seen": 11535176,
      "step": 17640
    },
    {
      "epoch": 9.2479035639413,
      "grad_norm": 0.4803924858570099,
      "learning_rate": 3.2528549272283016e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 11539016,
      "step": 17645
    },
    {
      "epoch": 9.250524109014675,
      "grad_norm": 0.3523067235946655,
      "learning_rate": 3.251764498760683e-05,
      "loss": 0.3776,
      "num_input_tokens_seen": 11543752,
      "step": 17650
    },
    {
      "epoch": 9.25314465408805,
      "grad_norm": 0.6476536393165588,
      "learning_rate": 3.2506739130325084e-05,
      "loss": 0.4506,
      "num_input_tokens_seen": 11546664,
      "step": 17655
    },
    {
      "epoch": 9.255765199161425,
      "grad_norm": 0.7563745379447937,
      "learning_rate": 3.2495831702719146e-05,
      "loss": 0.4762,
      "num_input_tokens_seen": 11549032,
      "step": 17660
    },
    {
      "epoch": 9.2583857442348,
      "grad_norm": 0.4392692744731903,
      "learning_rate": 3.248492270707074e-05,
      "loss": 0.5587,
      "num_input_tokens_seen": 11552328,
      "step": 17665
    },
    {
      "epoch": 9.261006289308176,
      "grad_norm": 0.4047268331050873,
      "learning_rate": 3.247401214566189e-05,
      "loss": 0.5597,
      "num_input_tokens_seen": 11555464,
      "step": 17670
    },
    {
      "epoch": 9.26362683438155,
      "grad_norm": 0.5589735507965088,
      "learning_rate": 3.2463100020774965e-05,
      "loss": 0.462,
      "num_input_tokens_seen": 11558088,
      "step": 17675
    },
    {
      "epoch": 9.266247379454926,
      "grad_norm": 0.4364038109779358,
      "learning_rate": 3.245218633469267e-05,
      "loss": 0.4988,
      "num_input_tokens_seen": 11560872,
      "step": 17680
    },
    {
      "epoch": 9.268867924528301,
      "grad_norm": 0.5030288696289062,
      "learning_rate": 3.244127108969801e-05,
      "loss": 0.405,
      "num_input_tokens_seen": 11564264,
      "step": 17685
    },
    {
      "epoch": 9.271488469601676,
      "grad_norm": 0.3574214577674866,
      "learning_rate": 3.2430354288074314e-05,
      "loss": 0.422,
      "num_input_tokens_seen": 11567656,
      "step": 17690
    },
    {
      "epoch": 9.274109014675052,
      "grad_norm": 0.41441887617111206,
      "learning_rate": 3.241943593210529e-05,
      "loss": 0.4878,
      "num_input_tokens_seen": 11571176,
      "step": 17695
    },
    {
      "epoch": 9.276729559748428,
      "grad_norm": 0.37752920389175415,
      "learning_rate": 3.2408516024074916e-05,
      "loss": 0.5199,
      "num_input_tokens_seen": 11574216,
      "step": 17700
    },
    {
      "epoch": 9.279350104821804,
      "grad_norm": 0.5319563746452332,
      "learning_rate": 3.2397594566267505e-05,
      "loss": 0.3569,
      "num_input_tokens_seen": 11577608,
      "step": 17705
    },
    {
      "epoch": 9.281970649895179,
      "grad_norm": 0.4268968105316162,
      "learning_rate": 3.23866715609677e-05,
      "loss": 0.4835,
      "num_input_tokens_seen": 11581000,
      "step": 17710
    },
    {
      "epoch": 9.284591194968554,
      "grad_norm": 0.47811368107795715,
      "learning_rate": 3.237574701046049e-05,
      "loss": 0.4723,
      "num_input_tokens_seen": 11584840,
      "step": 17715
    },
    {
      "epoch": 9.28721174004193,
      "grad_norm": 0.5160954594612122,
      "learning_rate": 3.236482091703115e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 11587912,
      "step": 17720
    },
    {
      "epoch": 9.289832285115304,
      "grad_norm": 0.6525090336799622,
      "learning_rate": 3.235389328296528e-05,
      "loss": 0.4686,
      "num_input_tokens_seen": 11591752,
      "step": 17725
    },
    {
      "epoch": 9.29245283018868,
      "grad_norm": 0.6126468777656555,
      "learning_rate": 3.2342964110548845e-05,
      "loss": 0.5239,
      "num_input_tokens_seen": 11594248,
      "step": 17730
    },
    {
      "epoch": 9.295073375262055,
      "grad_norm": 0.452462762594223,
      "learning_rate": 3.2332033402068093e-05,
      "loss": 0.447,
      "num_input_tokens_seen": 11597448,
      "step": 17735
    },
    {
      "epoch": 9.29769392033543,
      "grad_norm": 0.49328485131263733,
      "learning_rate": 3.2321101159809605e-05,
      "loss": 0.6286,
      "num_input_tokens_seen": 11600360,
      "step": 17740
    },
    {
      "epoch": 9.300314465408805,
      "grad_norm": 0.6699115037918091,
      "learning_rate": 3.231016738606027e-05,
      "loss": 0.5121,
      "num_input_tokens_seen": 11603528,
      "step": 17745
    },
    {
      "epoch": 9.30293501048218,
      "grad_norm": 0.5340557098388672,
      "learning_rate": 3.2299232083107313e-05,
      "loss": 0.486,
      "num_input_tokens_seen": 11606600,
      "step": 17750
    },
    {
      "epoch": 9.305555555555555,
      "grad_norm": 0.490557998418808,
      "learning_rate": 3.228829525323829e-05,
      "loss": 0.4143,
      "num_input_tokens_seen": 11609064,
      "step": 17755
    },
    {
      "epoch": 9.30817610062893,
      "grad_norm": 0.3499566912651062,
      "learning_rate": 3.227735689874104e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 11612328,
      "step": 17760
    },
    {
      "epoch": 9.310796645702306,
      "grad_norm": 0.39495113492012024,
      "learning_rate": 3.226641702190375e-05,
      "loss": 0.4663,
      "num_input_tokens_seen": 11615912,
      "step": 17765
    },
    {
      "epoch": 9.31341719077568,
      "grad_norm": 0.4942700266838074,
      "learning_rate": 3.225547562501493e-05,
      "loss": 0.496,
      "num_input_tokens_seen": 11618888,
      "step": 17770
    },
    {
      "epoch": 9.316037735849056,
      "grad_norm": 0.40651199221611023,
      "learning_rate": 3.2244532710363375e-05,
      "loss": 0.6834,
      "num_input_tokens_seen": 11621544,
      "step": 17775
    },
    {
      "epoch": 9.318658280922431,
      "grad_norm": 0.5088180899620056,
      "learning_rate": 3.2233588280238224e-05,
      "loss": 0.3979,
      "num_input_tokens_seen": 11624040,
      "step": 17780
    },
    {
      "epoch": 9.321278825995806,
      "grad_norm": 0.5111942291259766,
      "learning_rate": 3.2222642336928944e-05,
      "loss": 0.4633,
      "num_input_tokens_seen": 11626728,
      "step": 17785
    },
    {
      "epoch": 9.323899371069182,
      "grad_norm": 0.6280831694602966,
      "learning_rate": 3.2211694882725275e-05,
      "loss": 0.3383,
      "num_input_tokens_seen": 11629832,
      "step": 17790
    },
    {
      "epoch": 9.326519916142558,
      "grad_norm": 0.43891268968582153,
      "learning_rate": 3.220074591991731e-05,
      "loss": 0.3501,
      "num_input_tokens_seen": 11633320,
      "step": 17795
    },
    {
      "epoch": 9.329140461215934,
      "grad_norm": 0.41319844126701355,
      "learning_rate": 3.218979545079546e-05,
      "loss": 0.4176,
      "num_input_tokens_seen": 11636104,
      "step": 17800
    },
    {
      "epoch": 9.331761006289309,
      "grad_norm": 0.3732883036136627,
      "learning_rate": 3.217884347765042e-05,
      "loss": 0.4103,
      "num_input_tokens_seen": 11639752,
      "step": 17805
    },
    {
      "epoch": 9.334381551362684,
      "grad_norm": 0.7802413105964661,
      "learning_rate": 3.216789000277323e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 11642664,
      "step": 17810
    },
    {
      "epoch": 9.33700209643606,
      "grad_norm": 0.46688181161880493,
      "learning_rate": 3.2156935028455224e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 11646024,
      "step": 17815
    },
    {
      "epoch": 9.339622641509434,
      "grad_norm": 0.35379350185394287,
      "learning_rate": 3.2145978556988056e-05,
      "loss": 0.4379,
      "num_input_tokens_seen": 11649192,
      "step": 17820
    },
    {
      "epoch": 9.34224318658281,
      "grad_norm": 0.31517109274864197,
      "learning_rate": 3.213502059066371e-05,
      "loss": 0.554,
      "num_input_tokens_seen": 11653192,
      "step": 17825
    },
    {
      "epoch": 9.344863731656185,
      "grad_norm": 0.9371067881584167,
      "learning_rate": 3.212406113177445e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 11657224,
      "step": 17830
    },
    {
      "epoch": 9.34748427672956,
      "grad_norm": 0.47134363651275635,
      "learning_rate": 3.211310018261287e-05,
      "loss": 0.5806,
      "num_input_tokens_seen": 11660328,
      "step": 17835
    },
    {
      "epoch": 9.350104821802935,
      "grad_norm": 0.39769911766052246,
      "learning_rate": 3.21021377454719e-05,
      "loss": 0.5001,
      "num_input_tokens_seen": 11663400,
      "step": 17840
    },
    {
      "epoch": 9.35272536687631,
      "grad_norm": 0.5204943418502808,
      "learning_rate": 3.209117382264473e-05,
      "loss": 0.4946,
      "num_input_tokens_seen": 11665992,
      "step": 17845
    },
    {
      "epoch": 9.355345911949685,
      "grad_norm": 0.7662177085876465,
      "learning_rate": 3.208020841642489e-05,
      "loss": 0.4085,
      "num_input_tokens_seen": 11669000,
      "step": 17850
    },
    {
      "epoch": 9.35796645702306,
      "grad_norm": 0.5850039124488831,
      "learning_rate": 3.206924152910623e-05,
      "loss": 0.4379,
      "num_input_tokens_seen": 11671976,
      "step": 17855
    },
    {
      "epoch": 9.360587002096436,
      "grad_norm": 0.3981678783893585,
      "learning_rate": 3.205827316298289e-05,
      "loss": 0.391,
      "num_input_tokens_seen": 11674984,
      "step": 17860
    },
    {
      "epoch": 9.36320754716981,
      "grad_norm": 0.570095419883728,
      "learning_rate": 3.204730332034934e-05,
      "loss": 0.4251,
      "num_input_tokens_seen": 11677864,
      "step": 17865
    },
    {
      "epoch": 9.365828092243186,
      "grad_norm": 0.2069791704416275,
      "learning_rate": 3.2036332003500324e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 11681128,
      "step": 17870
    },
    {
      "epoch": 9.368448637316561,
      "grad_norm": 0.6043143272399902,
      "learning_rate": 3.2025359214730924e-05,
      "loss": 0.4518,
      "num_input_tokens_seen": 11683880,
      "step": 17875
    },
    {
      "epoch": 9.371069182389936,
      "grad_norm": 0.780064582824707,
      "learning_rate": 3.201438495633654e-05,
      "loss": 0.4585,
      "num_input_tokens_seen": 11687496,
      "step": 17880
    },
    {
      "epoch": 9.373689727463312,
      "grad_norm": 0.6530006527900696,
      "learning_rate": 3.200340923061284e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 11690696,
      "step": 17885
    },
    {
      "epoch": 9.376310272536688,
      "grad_norm": 0.362638920545578,
      "learning_rate": 3.199243203985583e-05,
      "loss": 0.4091,
      "num_input_tokens_seen": 11693544,
      "step": 17890
    },
    {
      "epoch": 9.378930817610064,
      "grad_norm": 0.5461180210113525,
      "learning_rate": 3.1981453386361804e-05,
      "loss": 0.5181,
      "num_input_tokens_seen": 11696840,
      "step": 17895
    },
    {
      "epoch": 9.381551362683439,
      "grad_norm": 0.3329000174999237,
      "learning_rate": 3.1970473272427386e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 11700264,
      "step": 17900
    },
    {
      "epoch": 9.384171907756814,
      "grad_norm": 0.6603819727897644,
      "learning_rate": 3.195949170034948e-05,
      "loss": 0.3216,
      "num_input_tokens_seen": 11703496,
      "step": 17905
    },
    {
      "epoch": 9.38679245283019,
      "grad_norm": 0.5842118263244629,
      "learning_rate": 3.19485086724253e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 11706664,
      "step": 17910
    },
    {
      "epoch": 9.389412997903564,
      "grad_norm": 0.6684367060661316,
      "learning_rate": 3.193752419095239e-05,
      "loss": 0.4793,
      "num_input_tokens_seen": 11709480,
      "step": 17915
    },
    {
      "epoch": 9.39203354297694,
      "grad_norm": 0.32089510560035706,
      "learning_rate": 3.192653825822857e-05,
      "loss": 0.3703,
      "num_input_tokens_seen": 11712712,
      "step": 17920
    },
    {
      "epoch": 9.394654088050315,
      "grad_norm": 0.6549530625343323,
      "learning_rate": 3.1915550876551964e-05,
      "loss": 0.5531,
      "num_input_tokens_seen": 11715464,
      "step": 17925
    },
    {
      "epoch": 9.39727463312369,
      "grad_norm": 0.5345918536186218,
      "learning_rate": 3.190456204822101e-05,
      "loss": 0.4971,
      "num_input_tokens_seen": 11718120,
      "step": 17930
    },
    {
      "epoch": 9.399895178197065,
      "grad_norm": 0.47410258650779724,
      "learning_rate": 3.1893571775534445e-05,
      "loss": 0.4956,
      "num_input_tokens_seen": 11721512,
      "step": 17935
    },
    {
      "epoch": 9.40251572327044,
      "grad_norm": 0.4812454283237457,
      "learning_rate": 3.1882580060791315e-05,
      "loss": 0.3104,
      "num_input_tokens_seen": 11724008,
      "step": 17940
    },
    {
      "epoch": 9.405136268343815,
      "grad_norm": 0.47444379329681396,
      "learning_rate": 3.187158690629095e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 11726696,
      "step": 17945
    },
    {
      "epoch": 9.40775681341719,
      "grad_norm": 0.24900157749652863,
      "learning_rate": 3.1860592314333e-05,
      "loss": 0.3676,
      "num_input_tokens_seen": 11730408,
      "step": 17950
    },
    {
      "epoch": 9.410377358490566,
      "grad_norm": 0.5313357710838318,
      "learning_rate": 3.184959628721742e-05,
      "loss": 0.5394,
      "num_input_tokens_seen": 11733320,
      "step": 17955
    },
    {
      "epoch": 9.41299790356394,
      "grad_norm": 0.6880872845649719,
      "learning_rate": 3.183859882724443e-05,
      "loss": 0.4404,
      "num_input_tokens_seen": 11738472,
      "step": 17960
    },
    {
      "epoch": 9.415618448637316,
      "grad_norm": 0.4520416855812073,
      "learning_rate": 3.1827599936714584e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 11741032,
      "step": 17965
    },
    {
      "epoch": 9.418238993710691,
      "grad_norm": 0.8593448996543884,
      "learning_rate": 3.1816599617928736e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 11744328,
      "step": 17970
    },
    {
      "epoch": 9.420859538784066,
      "grad_norm": 0.5697822570800781,
      "learning_rate": 3.1805597873188e-05,
      "loss": 0.3585,
      "num_input_tokens_seen": 11747976,
      "step": 17975
    },
    {
      "epoch": 9.423480083857442,
      "grad_norm": 0.36781617999076843,
      "learning_rate": 3.179459470479384e-05,
      "loss": 0.5602,
      "num_input_tokens_seen": 11751144,
      "step": 17980
    },
    {
      "epoch": 9.426100628930818,
      "grad_norm": 0.3811553716659546,
      "learning_rate": 3.1783590115047983e-05,
      "loss": 0.4803,
      "num_input_tokens_seen": 11754312,
      "step": 17985
    },
    {
      "epoch": 9.428721174004194,
      "grad_norm": 0.354647159576416,
      "learning_rate": 3.177258410625246e-05,
      "loss": 0.3448,
      "num_input_tokens_seen": 11757896,
      "step": 17990
    },
    {
      "epoch": 9.431341719077569,
      "grad_norm": 0.3234236240386963,
      "learning_rate": 3.176157668070962e-05,
      "loss": 0.3938,
      "num_input_tokens_seen": 11760840,
      "step": 17995
    },
    {
      "epoch": 9.433962264150944,
      "grad_norm": 0.48271381855010986,
      "learning_rate": 3.175056784072207e-05,
      "loss": 0.3992,
      "num_input_tokens_seen": 11764136,
      "step": 18000
    },
    {
      "epoch": 9.43658280922432,
      "grad_norm": 1.458939790725708,
      "learning_rate": 3.173955758859274e-05,
      "loss": 0.5595,
      "num_input_tokens_seen": 11766696,
      "step": 18005
    },
    {
      "epoch": 9.439203354297694,
      "grad_norm": 0.6974996328353882,
      "learning_rate": 3.172854592662486e-05,
      "loss": 0.5049,
      "num_input_tokens_seen": 11769160,
      "step": 18010
    },
    {
      "epoch": 9.44182389937107,
      "grad_norm": 0.2714075446128845,
      "learning_rate": 3.1717532857121926e-05,
      "loss": 0.4848,
      "num_input_tokens_seen": 11772360,
      "step": 18015
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 0.6915579438209534,
      "learning_rate": 3.170651838238776e-05,
      "loss": 0.51,
      "num_input_tokens_seen": 11774984,
      "step": 18020
    },
    {
      "epoch": 9.44706498951782,
      "grad_norm": 0.27692902088165283,
      "learning_rate": 3.169550250472646e-05,
      "loss": 0.4418,
      "num_input_tokens_seen": 11777928,
      "step": 18025
    },
    {
      "epoch": 9.449685534591195,
      "grad_norm": 0.43934178352355957,
      "learning_rate": 3.168448522644243e-05,
      "loss": 0.4621,
      "num_input_tokens_seen": 11781096,
      "step": 18030
    },
    {
      "epoch": 9.45230607966457,
      "grad_norm": 0.29469186067581177,
      "learning_rate": 3.167346654984034e-05,
      "loss": 0.3992,
      "num_input_tokens_seen": 11784104,
      "step": 18035
    },
    {
      "epoch": 9.454926624737945,
      "grad_norm": 0.2741835415363312,
      "learning_rate": 3.166244647722519e-05,
      "loss": 0.4832,
      "num_input_tokens_seen": 11787304,
      "step": 18040
    },
    {
      "epoch": 9.45754716981132,
      "grad_norm": 0.5683513879776001,
      "learning_rate": 3.1651425010902244e-05,
      "loss": 0.3471,
      "num_input_tokens_seen": 11790024,
      "step": 18045
    },
    {
      "epoch": 9.460167714884696,
      "grad_norm": 0.4325700104236603,
      "learning_rate": 3.164040215317706e-05,
      "loss": 0.3896,
      "num_input_tokens_seen": 11792840,
      "step": 18050
    },
    {
      "epoch": 9.46278825995807,
      "grad_norm": 0.6381353735923767,
      "learning_rate": 3.162937790635552e-05,
      "loss": 0.4303,
      "num_input_tokens_seen": 11795720,
      "step": 18055
    },
    {
      "epoch": 9.465408805031446,
      "grad_norm": 0.6165827512741089,
      "learning_rate": 3.1618352272743746e-05,
      "loss": 0.4348,
      "num_input_tokens_seen": 11798472,
      "step": 18060
    },
    {
      "epoch": 9.468029350104821,
      "grad_norm": 0.46312737464904785,
      "learning_rate": 3.160732525464818e-05,
      "loss": 0.4319,
      "num_input_tokens_seen": 11801800,
      "step": 18065
    },
    {
      "epoch": 9.470649895178196,
      "grad_norm": 0.6194776296615601,
      "learning_rate": 3.1596296854375554e-05,
      "loss": 0.5155,
      "num_input_tokens_seen": 11805160,
      "step": 18070
    },
    {
      "epoch": 9.473270440251572,
      "grad_norm": 0.5877133011817932,
      "learning_rate": 3.1585267074232884e-05,
      "loss": 0.4245,
      "num_input_tokens_seen": 11811816,
      "step": 18075
    },
    {
      "epoch": 9.475890985324948,
      "grad_norm": 0.24974000453948975,
      "learning_rate": 3.157423591652746e-05,
      "loss": 0.4962,
      "num_input_tokens_seen": 11816040,
      "step": 18080
    },
    {
      "epoch": 9.478511530398324,
      "grad_norm": 0.36367639899253845,
      "learning_rate": 3.1563203383566894e-05,
      "loss": 0.493,
      "num_input_tokens_seen": 11818920,
      "step": 18085
    },
    {
      "epoch": 9.481132075471699,
      "grad_norm": 0.6555071473121643,
      "learning_rate": 3.155216947765906e-05,
      "loss": 0.5176,
      "num_input_tokens_seen": 11821736,
      "step": 18090
    },
    {
      "epoch": 9.483752620545074,
      "grad_norm": 0.9384156465530396,
      "learning_rate": 3.154113420111211e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 11825256,
      "step": 18095
    },
    {
      "epoch": 9.48637316561845,
      "grad_norm": 0.2645401060581207,
      "learning_rate": 3.1530097556234526e-05,
      "loss": 0.5519,
      "num_input_tokens_seen": 11829416,
      "step": 18100
    },
    {
      "epoch": 9.488993710691824,
      "grad_norm": 0.3572096824645996,
      "learning_rate": 3.151905954533502e-05,
      "loss": 0.5679,
      "num_input_tokens_seen": 11832200,
      "step": 18105
    },
    {
      "epoch": 9.4916142557652,
      "grad_norm": 0.5735821723937988,
      "learning_rate": 3.150802017072263e-05,
      "loss": 0.4741,
      "num_input_tokens_seen": 11835592,
      "step": 18110
    },
    {
      "epoch": 9.494234800838575,
      "grad_norm": 0.4697646498680115,
      "learning_rate": 3.149697943470668e-05,
      "loss": 0.4168,
      "num_input_tokens_seen": 11839400,
      "step": 18115
    },
    {
      "epoch": 9.49685534591195,
      "grad_norm": 0.5356315970420837,
      "learning_rate": 3.1485937339596736e-05,
      "loss": 0.5559,
      "num_input_tokens_seen": 11842216,
      "step": 18120
    },
    {
      "epoch": 9.499475890985325,
      "grad_norm": 0.5049760341644287,
      "learning_rate": 3.147489388770271e-05,
      "loss": 0.3745,
      "num_input_tokens_seen": 11844936,
      "step": 18125
    },
    {
      "epoch": 9.5020964360587,
      "grad_norm": 0.6743546724319458,
      "learning_rate": 3.146384908133475e-05,
      "loss": 0.4576,
      "num_input_tokens_seen": 11847912,
      "step": 18130
    },
    {
      "epoch": 9.504716981132075,
      "grad_norm": 0.24800238013267517,
      "learning_rate": 3.14528029228033e-05,
      "loss": 0.3561,
      "num_input_tokens_seen": 11853192,
      "step": 18135
    },
    {
      "epoch": 9.50733752620545,
      "grad_norm": 0.6463057398796082,
      "learning_rate": 3.144175541441911e-05,
      "loss": 0.5263,
      "num_input_tokens_seen": 11856808,
      "step": 18140
    },
    {
      "epoch": 9.509958071278826,
      "grad_norm": 0.7637888789176941,
      "learning_rate": 3.1430706558493166e-05,
      "loss": 0.6983,
      "num_input_tokens_seen": 11860360,
      "step": 18145
    },
    {
      "epoch": 9.5125786163522,
      "grad_norm": 0.42317572236061096,
      "learning_rate": 3.141965635733677e-05,
      "loss": 0.3473,
      "num_input_tokens_seen": 11863048,
      "step": 18150
    },
    {
      "epoch": 9.515199161425576,
      "grad_norm": 0.3443625271320343,
      "learning_rate": 3.140860481326152e-05,
      "loss": 0.4294,
      "num_input_tokens_seen": 11867016,
      "step": 18155
    },
    {
      "epoch": 9.517819706498951,
      "grad_norm": 0.73611980676651,
      "learning_rate": 3.139755192857925e-05,
      "loss": 0.4272,
      "num_input_tokens_seen": 11870728,
      "step": 18160
    },
    {
      "epoch": 9.520440251572326,
      "grad_norm": 0.3176687955856323,
      "learning_rate": 3.1386497705602094e-05,
      "loss": 0.4713,
      "num_input_tokens_seen": 11873928,
      "step": 18165
    },
    {
      "epoch": 9.523060796645701,
      "grad_norm": 0.513369619846344,
      "learning_rate": 3.137544214664249e-05,
      "loss": 0.4031,
      "num_input_tokens_seen": 11876968,
      "step": 18170
    },
    {
      "epoch": 9.525681341719078,
      "grad_norm": 0.3259904682636261,
      "learning_rate": 3.136438525401312e-05,
      "loss": 0.4412,
      "num_input_tokens_seen": 11880648,
      "step": 18175
    },
    {
      "epoch": 9.528301886792454,
      "grad_norm": 0.5462003350257874,
      "learning_rate": 3.135332703002696e-05,
      "loss": 0.4526,
      "num_input_tokens_seen": 11883560,
      "step": 18180
    },
    {
      "epoch": 9.530922431865829,
      "grad_norm": 0.4204510748386383,
      "learning_rate": 3.1342267476997276e-05,
      "loss": 0.4932,
      "num_input_tokens_seen": 11888296,
      "step": 18185
    },
    {
      "epoch": 9.533542976939204,
      "grad_norm": 0.3847810626029968,
      "learning_rate": 3.1331206597237584e-05,
      "loss": 0.4171,
      "num_input_tokens_seen": 11892872,
      "step": 18190
    },
    {
      "epoch": 9.536163522012579,
      "grad_norm": 0.7511320114135742,
      "learning_rate": 3.1320144393061715e-05,
      "loss": 0.3999,
      "num_input_tokens_seen": 11895944,
      "step": 18195
    },
    {
      "epoch": 9.538784067085954,
      "grad_norm": 0.8128964304924011,
      "learning_rate": 3.130908086678373e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 11898696,
      "step": 18200
    },
    {
      "epoch": 9.54140461215933,
      "grad_norm": 0.27898985147476196,
      "learning_rate": 3.1298016020718014e-05,
      "loss": 0.3494,
      "num_input_tokens_seen": 11904744,
      "step": 18205
    },
    {
      "epoch": 9.544025157232705,
      "grad_norm": 0.6567946672439575,
      "learning_rate": 3.128694985717919e-05,
      "loss": 0.4173,
      "num_input_tokens_seen": 11907656,
      "step": 18210
    },
    {
      "epoch": 9.54664570230608,
      "grad_norm": 0.463409423828125,
      "learning_rate": 3.1275882378482185e-05,
      "loss": 0.5304,
      "num_input_tokens_seen": 11910504,
      "step": 18215
    },
    {
      "epoch": 9.549266247379455,
      "grad_norm": 1.2063987255096436,
      "learning_rate": 3.1264813586942185e-05,
      "loss": 0.5248,
      "num_input_tokens_seen": 11913192,
      "step": 18220
    },
    {
      "epoch": 9.55188679245283,
      "grad_norm": 0.4877212345600128,
      "learning_rate": 3.125374348487466e-05,
      "loss": 0.5911,
      "num_input_tokens_seen": 11916680,
      "step": 18225
    },
    {
      "epoch": 9.554507337526205,
      "grad_norm": 1.0774281024932861,
      "learning_rate": 3.1242672074595334e-05,
      "loss": 0.3972,
      "num_input_tokens_seen": 11919656,
      "step": 18230
    },
    {
      "epoch": 9.55712788259958,
      "grad_norm": 0.37588292360305786,
      "learning_rate": 3.123159935842024e-05,
      "loss": 0.4188,
      "num_input_tokens_seen": 11922632,
      "step": 18235
    },
    {
      "epoch": 9.559748427672956,
      "grad_norm": 0.4620734751224518,
      "learning_rate": 3.122052533866565e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 11926280,
      "step": 18240
    },
    {
      "epoch": 9.56236897274633,
      "grad_norm": 0.3171112835407257,
      "learning_rate": 3.120945001764813e-05,
      "loss": 0.3018,
      "num_input_tokens_seen": 11929768,
      "step": 18245
    },
    {
      "epoch": 9.564989517819706,
      "grad_norm": 0.32013556361198425,
      "learning_rate": 3.11983733976845e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 11933480,
      "step": 18250
    },
    {
      "epoch": 9.567610062893081,
      "grad_norm": 0.4293838143348694,
      "learning_rate": 3.1187295481091876e-05,
      "loss": 0.3496,
      "num_input_tokens_seen": 11936200,
      "step": 18255
    },
    {
      "epoch": 9.570230607966456,
      "grad_norm": 0.5979850888252258,
      "learning_rate": 3.1176216270187625e-05,
      "loss": 0.545,
      "num_input_tokens_seen": 11938472,
      "step": 18260
    },
    {
      "epoch": 9.572851153039831,
      "grad_norm": 0.29470670223236084,
      "learning_rate": 3.116513576728939e-05,
      "loss": 0.5276,
      "num_input_tokens_seen": 11942280,
      "step": 18265
    },
    {
      "epoch": 9.575471698113208,
      "grad_norm": 0.3512853980064392,
      "learning_rate": 3.1154053974715085e-05,
      "loss": 0.4063,
      "num_input_tokens_seen": 11945800,
      "step": 18270
    },
    {
      "epoch": 9.578092243186584,
      "grad_norm": 0.32676130533218384,
      "learning_rate": 3.1142970894782906e-05,
      "loss": 0.4206,
      "num_input_tokens_seen": 11948200,
      "step": 18275
    },
    {
      "epoch": 9.580712788259959,
      "grad_norm": 0.48089778423309326,
      "learning_rate": 3.1131886529811296e-05,
      "loss": 0.4862,
      "num_input_tokens_seen": 11951656,
      "step": 18280
    },
    {
      "epoch": 9.583333333333334,
      "grad_norm": 0.31764474511146545,
      "learning_rate": 3.1120800882118985e-05,
      "loss": 0.4459,
      "num_input_tokens_seen": 11955208,
      "step": 18285
    },
    {
      "epoch": 9.585953878406709,
      "grad_norm": 0.4673563539981842,
      "learning_rate": 3.1109713954024945e-05,
      "loss": 0.3631,
      "num_input_tokens_seen": 11957768,
      "step": 18290
    },
    {
      "epoch": 9.588574423480084,
      "grad_norm": 0.5650528073310852,
      "learning_rate": 3.109862574784846e-05,
      "loss": 0.4374,
      "num_input_tokens_seen": 11960552,
      "step": 18295
    },
    {
      "epoch": 9.59119496855346,
      "grad_norm": 0.5599598288536072,
      "learning_rate": 3.108753626590905e-05,
      "loss": 0.5144,
      "num_input_tokens_seen": 11964552,
      "step": 18300
    },
    {
      "epoch": 9.593815513626835,
      "grad_norm": 0.9603380560874939,
      "learning_rate": 3.1076445510526486e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 11966824,
      "step": 18305
    },
    {
      "epoch": 9.59643605870021,
      "grad_norm": 0.5078924894332886,
      "learning_rate": 3.106535348402086e-05,
      "loss": 0.4368,
      "num_input_tokens_seen": 11969832,
      "step": 18310
    },
    {
      "epoch": 9.599056603773585,
      "grad_norm": 0.3684995770454407,
      "learning_rate": 3.105426018871247e-05,
      "loss": 0.3894,
      "num_input_tokens_seen": 11972424,
      "step": 18315
    },
    {
      "epoch": 9.60167714884696,
      "grad_norm": 0.5410547852516174,
      "learning_rate": 3.104316562692193e-05,
      "loss": 0.4654,
      "num_input_tokens_seen": 11976552,
      "step": 18320
    },
    {
      "epoch": 9.604297693920335,
      "grad_norm": 0.49314776062965393,
      "learning_rate": 3.103206980097009e-05,
      "loss": 0.427,
      "num_input_tokens_seen": 11979720,
      "step": 18325
    },
    {
      "epoch": 9.60691823899371,
      "grad_norm": 0.37157875299453735,
      "learning_rate": 3.102097271317805e-05,
      "loss": 0.5186,
      "num_input_tokens_seen": 11982536,
      "step": 18330
    },
    {
      "epoch": 9.609538784067086,
      "grad_norm": 0.30451497435569763,
      "learning_rate": 3.100987436586721e-05,
      "loss": 0.3311,
      "num_input_tokens_seen": 11985384,
      "step": 18335
    },
    {
      "epoch": 9.61215932914046,
      "grad_norm": 0.44640833139419556,
      "learning_rate": 3.099877476135923e-05,
      "loss": 0.409,
      "num_input_tokens_seen": 11988712,
      "step": 18340
    },
    {
      "epoch": 9.614779874213836,
      "grad_norm": 0.2395281046628952,
      "learning_rate": 3.0987673901976e-05,
      "loss": 0.4226,
      "num_input_tokens_seen": 11993352,
      "step": 18345
    },
    {
      "epoch": 9.617400419287211,
      "grad_norm": 0.33634066581726074,
      "learning_rate": 3.0976571790039697e-05,
      "loss": 0.4582,
      "num_input_tokens_seen": 11995848,
      "step": 18350
    },
    {
      "epoch": 9.620020964360586,
      "grad_norm": 0.4215756952762604,
      "learning_rate": 3.096546842787277e-05,
      "loss": 0.5118,
      "num_input_tokens_seen": 11999336,
      "step": 18355
    },
    {
      "epoch": 9.622641509433961,
      "grad_norm": 0.480035662651062,
      "learning_rate": 3.095436381779789e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 12003080,
      "step": 18360
    },
    {
      "epoch": 9.625262054507338,
      "grad_norm": 0.28751227259635925,
      "learning_rate": 3.0943257962138045e-05,
      "loss": 0.2732,
      "num_input_tokens_seen": 12007240,
      "step": 18365
    },
    {
      "epoch": 9.627882599580714,
      "grad_norm": 0.38190287351608276,
      "learning_rate": 3.093215086321644e-05,
      "loss": 0.4904,
      "num_input_tokens_seen": 12009928,
      "step": 18370
    },
    {
      "epoch": 9.630503144654089,
      "grad_norm": 0.3371591567993164,
      "learning_rate": 3.092104252335654e-05,
      "loss": 0.4156,
      "num_input_tokens_seen": 12013384,
      "step": 18375
    },
    {
      "epoch": 9.633123689727464,
      "grad_norm": 0.6858401894569397,
      "learning_rate": 3.09099329448821e-05,
      "loss": 0.4839,
      "num_input_tokens_seen": 12016520,
      "step": 18380
    },
    {
      "epoch": 9.635744234800839,
      "grad_norm": 0.6989974975585938,
      "learning_rate": 3.089882213011711e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 12020040,
      "step": 18385
    },
    {
      "epoch": 9.638364779874214,
      "grad_norm": 0.4420827031135559,
      "learning_rate": 3.088771008138583e-05,
      "loss": 0.4365,
      "num_input_tokens_seen": 12023464,
      "step": 18390
    },
    {
      "epoch": 9.64098532494759,
      "grad_norm": 0.5866673588752747,
      "learning_rate": 3.087659680101278e-05,
      "loss": 0.3257,
      "num_input_tokens_seen": 12026664,
      "step": 18395
    },
    {
      "epoch": 9.643605870020965,
      "grad_norm": 0.3224429488182068,
      "learning_rate": 3.086548229132271e-05,
      "loss": 0.3866,
      "num_input_tokens_seen": 12030440,
      "step": 18400
    },
    {
      "epoch": 9.64622641509434,
      "grad_norm": 0.48861560225486755,
      "learning_rate": 3.085436655464067e-05,
      "loss": 0.4902,
      "num_input_tokens_seen": 12032872,
      "step": 18405
    },
    {
      "epoch": 9.648846960167715,
      "grad_norm": 0.401684045791626,
      "learning_rate": 3.084324959329193e-05,
      "loss": 0.3761,
      "num_input_tokens_seen": 12037864,
      "step": 18410
    },
    {
      "epoch": 9.65146750524109,
      "grad_norm": 0.4396824538707733,
      "learning_rate": 3.083213140960204e-05,
      "loss": 0.3021,
      "num_input_tokens_seen": 12040904,
      "step": 18415
    },
    {
      "epoch": 9.654088050314465,
      "grad_norm": 0.45827412605285645,
      "learning_rate": 3.08210120058968e-05,
      "loss": 0.4402,
      "num_input_tokens_seen": 12044296,
      "step": 18420
    },
    {
      "epoch": 9.65670859538784,
      "grad_norm": 0.39161089062690735,
      "learning_rate": 3.080989138450225e-05,
      "loss": 0.3459,
      "num_input_tokens_seen": 12047208,
      "step": 18425
    },
    {
      "epoch": 9.659329140461216,
      "grad_norm": 0.43163034319877625,
      "learning_rate": 3.0798769547744713e-05,
      "loss": 0.4966,
      "num_input_tokens_seen": 12050696,
      "step": 18430
    },
    {
      "epoch": 9.66194968553459,
      "grad_norm": 0.45074668526649475,
      "learning_rate": 3.0787646497950736e-05,
      "loss": 0.4903,
      "num_input_tokens_seen": 12054536,
      "step": 18435
    },
    {
      "epoch": 9.664570230607966,
      "grad_norm": 0.5946494340896606,
      "learning_rate": 3.0776522237447146e-05,
      "loss": 0.444,
      "num_input_tokens_seen": 12056776,
      "step": 18440
    },
    {
      "epoch": 9.667190775681341,
      "grad_norm": 0.5295713543891907,
      "learning_rate": 3.076539676856101e-05,
      "loss": 0.5012,
      "num_input_tokens_seen": 12059720,
      "step": 18445
    },
    {
      "epoch": 9.669811320754716,
      "grad_norm": 0.5394542217254639,
      "learning_rate": 3.0754270093619634e-05,
      "loss": 0.4409,
      "num_input_tokens_seen": 12063304,
      "step": 18450
    },
    {
      "epoch": 9.672431865828091,
      "grad_norm": 0.3979273736476898,
      "learning_rate": 3.074314221495061e-05,
      "loss": 0.5023,
      "num_input_tokens_seen": 12067592,
      "step": 18455
    },
    {
      "epoch": 9.675052410901468,
      "grad_norm": 0.39289161562919617,
      "learning_rate": 3.0732013134881756e-05,
      "loss": 0.457,
      "num_input_tokens_seen": 12070856,
      "step": 18460
    },
    {
      "epoch": 9.677672955974844,
      "grad_norm": 0.38029396533966064,
      "learning_rate": 3.072088285574114e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 12074280,
      "step": 18465
    },
    {
      "epoch": 9.680293501048219,
      "grad_norm": 0.4770740568637848,
      "learning_rate": 3.070975137985711e-05,
      "loss": 0.3966,
      "num_input_tokens_seen": 12077288,
      "step": 18470
    },
    {
      "epoch": 9.682914046121594,
      "grad_norm": 0.41529250144958496,
      "learning_rate": 3.0698618709558226e-05,
      "loss": 0.3722,
      "num_input_tokens_seen": 12082056,
      "step": 18475
    },
    {
      "epoch": 9.685534591194969,
      "grad_norm": 0.34995582699775696,
      "learning_rate": 3.068748484717332e-05,
      "loss": 0.5215,
      "num_input_tokens_seen": 12085128,
      "step": 18480
    },
    {
      "epoch": 9.688155136268344,
      "grad_norm": 0.3668736517429352,
      "learning_rate": 3.067634979503147e-05,
      "loss": 0.5416,
      "num_input_tokens_seen": 12089032,
      "step": 18485
    },
    {
      "epoch": 9.69077568134172,
      "grad_norm": 0.4269539713859558,
      "learning_rate": 3.0665213555462e-05,
      "loss": 0.5255,
      "num_input_tokens_seen": 12091496,
      "step": 18490
    },
    {
      "epoch": 9.693396226415095,
      "grad_norm": 0.42932167649269104,
      "learning_rate": 3.0654076130794494e-05,
      "loss": 0.5544,
      "num_input_tokens_seen": 12095208,
      "step": 18495
    },
    {
      "epoch": 9.69601677148847,
      "grad_norm": 0.4802780747413635,
      "learning_rate": 3.064293752335876e-05,
      "loss": 0.5079,
      "num_input_tokens_seen": 12099176,
      "step": 18500
    },
    {
      "epoch": 9.698637316561845,
      "grad_norm": 0.5171857476234436,
      "learning_rate": 3.063179773548487e-05,
      "loss": 0.5482,
      "num_input_tokens_seen": 12102408,
      "step": 18505
    },
    {
      "epoch": 9.70125786163522,
      "grad_norm": 0.43447375297546387,
      "learning_rate": 3.0620656769503164e-05,
      "loss": 0.537,
      "num_input_tokens_seen": 12105704,
      "step": 18510
    },
    {
      "epoch": 9.703878406708595,
      "grad_norm": 0.2664455771446228,
      "learning_rate": 3.060951462774417e-05,
      "loss": 0.5355,
      "num_input_tokens_seen": 12108520,
      "step": 18515
    },
    {
      "epoch": 9.70649895178197,
      "grad_norm": 0.401510089635849,
      "learning_rate": 3.0598371312538714e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 12111912,
      "step": 18520
    },
    {
      "epoch": 9.709119496855346,
      "grad_norm": 0.5735644102096558,
      "learning_rate": 3.058722682621785e-05,
      "loss": 0.4225,
      "num_input_tokens_seen": 12116040,
      "step": 18525
    },
    {
      "epoch": 9.71174004192872,
      "grad_norm": 0.3503003716468811,
      "learning_rate": 3.0576081171112875e-05,
      "loss": 0.405,
      "num_input_tokens_seen": 12118632,
      "step": 18530
    },
    {
      "epoch": 9.714360587002096,
      "grad_norm": 0.5797234177589417,
      "learning_rate": 3.056493434955534e-05,
      "loss": 0.3305,
      "num_input_tokens_seen": 12121192,
      "step": 18535
    },
    {
      "epoch": 9.716981132075471,
      "grad_norm": 0.5025907754898071,
      "learning_rate": 3.0553786363877027e-05,
      "loss": 0.4174,
      "num_input_tokens_seen": 12123976,
      "step": 18540
    },
    {
      "epoch": 9.719601677148846,
      "grad_norm": 0.37314626574516296,
      "learning_rate": 3.054263721640996e-05,
      "loss": 0.5667,
      "num_input_tokens_seen": 12127816,
      "step": 18545
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 0.3003219962120056,
      "learning_rate": 3.053148690948642e-05,
      "loss": 0.5044,
      "num_input_tokens_seen": 12131848,
      "step": 18550
    },
    {
      "epoch": 9.724842767295598,
      "grad_norm": 0.40588194131851196,
      "learning_rate": 3.0520335445438935e-05,
      "loss": 0.4234,
      "num_input_tokens_seen": 12135432,
      "step": 18555
    },
    {
      "epoch": 9.727463312368974,
      "grad_norm": 0.47960251569747925,
      "learning_rate": 3.0509182826600246e-05,
      "loss": 0.4842,
      "num_input_tokens_seen": 12138376,
      "step": 18560
    },
    {
      "epoch": 9.730083857442349,
      "grad_norm": 0.41008618474006653,
      "learning_rate": 3.049802905530337e-05,
      "loss": 0.5466,
      "num_input_tokens_seen": 12141480,
      "step": 18565
    },
    {
      "epoch": 9.732704402515724,
      "grad_norm": 0.5094690322875977,
      "learning_rate": 3.0486874133881533e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 12144168,
      "step": 18570
    },
    {
      "epoch": 9.735324947589099,
      "grad_norm": 0.37827879190444946,
      "learning_rate": 3.047571806466823e-05,
      "loss": 0.5628,
      "num_input_tokens_seen": 12148168,
      "step": 18575
    },
    {
      "epoch": 9.737945492662474,
      "grad_norm": 0.3977164030075073,
      "learning_rate": 3.0464560849997177e-05,
      "loss": 0.3219,
      "num_input_tokens_seen": 12150696,
      "step": 18580
    },
    {
      "epoch": 9.74056603773585,
      "grad_norm": 0.6264135837554932,
      "learning_rate": 3.0453402492202325e-05,
      "loss": 0.3535,
      "num_input_tokens_seen": 12153928,
      "step": 18585
    },
    {
      "epoch": 9.743186582809225,
      "grad_norm": 0.45421460270881653,
      "learning_rate": 3.0442242993617904e-05,
      "loss": 0.4168,
      "num_input_tokens_seen": 12157384,
      "step": 18590
    },
    {
      "epoch": 9.7458071278826,
      "grad_norm": 0.3547348380088806,
      "learning_rate": 3.0431082356578334e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 12160200,
      "step": 18595
    },
    {
      "epoch": 9.748427672955975,
      "grad_norm": 0.5576347708702087,
      "learning_rate": 3.0419920583418294e-05,
      "loss": 0.62,
      "num_input_tokens_seen": 12162504,
      "step": 18600
    },
    {
      "epoch": 9.75104821802935,
      "grad_norm": 0.49202853441238403,
      "learning_rate": 3.0408757676472694e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 12165288,
      "step": 18605
    },
    {
      "epoch": 9.753668763102725,
      "grad_norm": 0.8335981965065002,
      "learning_rate": 3.0397593638076705e-05,
      "loss": 0.4256,
      "num_input_tokens_seen": 12167912,
      "step": 18610
    },
    {
      "epoch": 9.7562893081761,
      "grad_norm": 1.872330904006958,
      "learning_rate": 3.0386428470565702e-05,
      "loss": 0.5285,
      "num_input_tokens_seen": 12171624,
      "step": 18615
    },
    {
      "epoch": 9.758909853249476,
      "grad_norm": 0.7955192923545837,
      "learning_rate": 3.0375262176275308e-05,
      "loss": 0.4715,
      "num_input_tokens_seen": 12174632,
      "step": 18620
    },
    {
      "epoch": 9.76153039832285,
      "grad_norm": 0.7235605120658875,
      "learning_rate": 3.03640947575414e-05,
      "loss": 0.5705,
      "num_input_tokens_seen": 12177896,
      "step": 18625
    },
    {
      "epoch": 9.764150943396226,
      "grad_norm": 0.4649966359138489,
      "learning_rate": 3.0352926216700067e-05,
      "loss": 0.4796,
      "num_input_tokens_seen": 12181256,
      "step": 18630
    },
    {
      "epoch": 9.766771488469601,
      "grad_norm": 0.43443894386291504,
      "learning_rate": 3.034175655608763e-05,
      "loss": 0.4967,
      "num_input_tokens_seen": 12183848,
      "step": 18635
    },
    {
      "epoch": 9.769392033542976,
      "grad_norm": 0.322812557220459,
      "learning_rate": 3.0330585778040677e-05,
      "loss": 0.4273,
      "num_input_tokens_seen": 12187208,
      "step": 18640
    },
    {
      "epoch": 9.772012578616351,
      "grad_norm": 0.29049718379974365,
      "learning_rate": 3.031941388489598e-05,
      "loss": 0.4684,
      "num_input_tokens_seen": 12190568,
      "step": 18645
    },
    {
      "epoch": 9.774633123689728,
      "grad_norm": 0.4711388051509857,
      "learning_rate": 3.0308240878990584e-05,
      "loss": 0.3231,
      "num_input_tokens_seen": 12193224,
      "step": 18650
    },
    {
      "epoch": 9.777253668763104,
      "grad_norm": 0.5717551708221436,
      "learning_rate": 3.0297066762661773e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 12195816,
      "step": 18655
    },
    {
      "epoch": 9.779874213836479,
      "grad_norm": 0.6811365485191345,
      "learning_rate": 3.0285891538247015e-05,
      "loss": 0.4513,
      "num_input_tokens_seen": 12198856,
      "step": 18660
    },
    {
      "epoch": 9.782494758909854,
      "grad_norm": 0.5612965822219849,
      "learning_rate": 3.0274715208084053e-05,
      "loss": 0.5318,
      "num_input_tokens_seen": 12201992,
      "step": 18665
    },
    {
      "epoch": 9.785115303983229,
      "grad_norm": 0.4614351987838745,
      "learning_rate": 3.0263537774510847e-05,
      "loss": 0.4632,
      "num_input_tokens_seen": 12205608,
      "step": 18670
    },
    {
      "epoch": 9.787735849056604,
      "grad_norm": 0.3507581055164337,
      "learning_rate": 3.0252359239865585e-05,
      "loss": 0.5238,
      "num_input_tokens_seen": 12209352,
      "step": 18675
    },
    {
      "epoch": 9.79035639412998,
      "grad_norm": 0.70916348695755,
      "learning_rate": 3.0241179606486704e-05,
      "loss": 0.4364,
      "num_input_tokens_seen": 12213768,
      "step": 18680
    },
    {
      "epoch": 9.792976939203355,
      "grad_norm": 0.6387481093406677,
      "learning_rate": 3.0229998876712827e-05,
      "loss": 0.4339,
      "num_input_tokens_seen": 12216776,
      "step": 18685
    },
    {
      "epoch": 9.79559748427673,
      "grad_norm": 0.3565674126148224,
      "learning_rate": 3.0218817052882858e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 12221000,
      "step": 18690
    },
    {
      "epoch": 9.798218029350105,
      "grad_norm": 0.7698875665664673,
      "learning_rate": 3.020763413733591e-05,
      "loss": 0.4524,
      "num_input_tokens_seen": 12224200,
      "step": 18695
    },
    {
      "epoch": 9.80083857442348,
      "grad_norm": 0.42391613125801086,
      "learning_rate": 3.0196450132411302e-05,
      "loss": 0.4582,
      "num_input_tokens_seen": 12226920,
      "step": 18700
    },
    {
      "epoch": 9.803459119496855,
      "grad_norm": 0.3391323685646057,
      "learning_rate": 3.0185265040448616e-05,
      "loss": 0.4926,
      "num_input_tokens_seen": 12229960,
      "step": 18705
    },
    {
      "epoch": 9.80607966457023,
      "grad_norm": 0.5565849542617798,
      "learning_rate": 3.017407886378764e-05,
      "loss": 0.4051,
      "num_input_tokens_seen": 12232040,
      "step": 18710
    },
    {
      "epoch": 9.808700209643606,
      "grad_norm": 0.42739632725715637,
      "learning_rate": 3.016289160476839e-05,
      "loss": 0.3848,
      "num_input_tokens_seen": 12234472,
      "step": 18715
    },
    {
      "epoch": 9.81132075471698,
      "grad_norm": 0.880944550037384,
      "learning_rate": 3.0151703265731125e-05,
      "loss": 0.3398,
      "num_input_tokens_seen": 12237448,
      "step": 18720
    },
    {
      "epoch": 9.813941299790356,
      "grad_norm": 0.33133068680763245,
      "learning_rate": 3.0140513849016305e-05,
      "loss": 0.4378,
      "num_input_tokens_seen": 12240264,
      "step": 18725
    },
    {
      "epoch": 9.816561844863731,
      "grad_norm": 0.40700891613960266,
      "learning_rate": 3.0129323356964627e-05,
      "loss": 0.3892,
      "num_input_tokens_seen": 12243176,
      "step": 18730
    },
    {
      "epoch": 9.819182389937106,
      "grad_norm": 0.5553548336029053,
      "learning_rate": 3.0118131791917032e-05,
      "loss": 0.4589,
      "num_input_tokens_seen": 12245864,
      "step": 18735
    },
    {
      "epoch": 9.821802935010481,
      "grad_norm": 0.41629737615585327,
      "learning_rate": 3.010693915621466e-05,
      "loss": 0.4291,
      "num_input_tokens_seen": 12248712,
      "step": 18740
    },
    {
      "epoch": 9.824423480083858,
      "grad_norm": 0.9911262392997742,
      "learning_rate": 3.009574545219887e-05,
      "loss": 0.5072,
      "num_input_tokens_seen": 12251144,
      "step": 18745
    },
    {
      "epoch": 9.827044025157234,
      "grad_norm": 0.5137784481048584,
      "learning_rate": 3.0084550682211272e-05,
      "loss": 0.4494,
      "num_input_tokens_seen": 12255336,
      "step": 18750
    },
    {
      "epoch": 9.829664570230609,
      "grad_norm": 0.7978218793869019,
      "learning_rate": 3.0073354848593677e-05,
      "loss": 0.4895,
      "num_input_tokens_seen": 12257672,
      "step": 18755
    },
    {
      "epoch": 9.832285115303984,
      "grad_norm": 0.3712886869907379,
      "learning_rate": 3.0062157953688125e-05,
      "loss": 0.4437,
      "num_input_tokens_seen": 12261032,
      "step": 18760
    },
    {
      "epoch": 9.834905660377359,
      "grad_norm": 0.49518439173698425,
      "learning_rate": 3.0050959999836887e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 12264488,
      "step": 18765
    },
    {
      "epoch": 9.837526205450734,
      "grad_norm": 0.6775668859481812,
      "learning_rate": 3.003976098938244e-05,
      "loss": 0.4564,
      "num_input_tokens_seen": 12267432,
      "step": 18770
    },
    {
      "epoch": 9.84014675052411,
      "grad_norm": 0.5630773901939392,
      "learning_rate": 3.0028560924667487e-05,
      "loss": 0.3365,
      "num_input_tokens_seen": 12270888,
      "step": 18775
    },
    {
      "epoch": 9.842767295597485,
      "grad_norm": 0.324101984500885,
      "learning_rate": 3.001735980803496e-05,
      "loss": 0.4326,
      "num_input_tokens_seen": 12273896,
      "step": 18780
    },
    {
      "epoch": 9.84538784067086,
      "grad_norm": 0.42324548959732056,
      "learning_rate": 3.0006157641828004e-05,
      "loss": 0.5012,
      "num_input_tokens_seen": 12276328,
      "step": 18785
    },
    {
      "epoch": 9.848008385744235,
      "grad_norm": 0.481418639421463,
      "learning_rate": 2.9994954428389977e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 12279016,
      "step": 18790
    },
    {
      "epoch": 9.85062893081761,
      "grad_norm": 0.6035082340240479,
      "learning_rate": 2.9983750170064472e-05,
      "loss": 0.3511,
      "num_input_tokens_seen": 12281512,
      "step": 18795
    },
    {
      "epoch": 9.853249475890985,
      "grad_norm": 0.5507065653800964,
      "learning_rate": 2.9972544869195292e-05,
      "loss": 0.5222,
      "num_input_tokens_seen": 12285160,
      "step": 18800
    },
    {
      "epoch": 9.85587002096436,
      "grad_norm": 0.4121074080467224,
      "learning_rate": 2.9961338528126447e-05,
      "loss": 0.4565,
      "num_input_tokens_seen": 12288616,
      "step": 18805
    },
    {
      "epoch": 9.858490566037736,
      "grad_norm": 0.3780803084373474,
      "learning_rate": 2.9950131149202192e-05,
      "loss": 0.5306,
      "num_input_tokens_seen": 12292680,
      "step": 18810
    },
    {
      "epoch": 9.86111111111111,
      "grad_norm": 0.389250785112381,
      "learning_rate": 2.9938922734766973e-05,
      "loss": 0.4019,
      "num_input_tokens_seen": 12297160,
      "step": 18815
    },
    {
      "epoch": 9.863731656184486,
      "grad_norm": 0.3786046802997589,
      "learning_rate": 2.992771328716546e-05,
      "loss": 0.5919,
      "num_input_tokens_seen": 12300520,
      "step": 18820
    },
    {
      "epoch": 9.866352201257861,
      "grad_norm": 0.4874766767024994,
      "learning_rate": 2.991650280874255e-05,
      "loss": 0.4558,
      "num_input_tokens_seen": 12303496,
      "step": 18825
    },
    {
      "epoch": 9.868972746331236,
      "grad_norm": 0.47765955328941345,
      "learning_rate": 2.990529130184334e-05,
      "loss": 0.5182,
      "num_input_tokens_seen": 12306568,
      "step": 18830
    },
    {
      "epoch": 9.871593291404611,
      "grad_norm": 0.43767204880714417,
      "learning_rate": 2.989407876881316e-05,
      "loss": 0.5361,
      "num_input_tokens_seen": 12309160,
      "step": 18835
    },
    {
      "epoch": 9.874213836477988,
      "grad_norm": 0.40697094798088074,
      "learning_rate": 2.988286521199753e-05,
      "loss": 0.4082,
      "num_input_tokens_seen": 12312776,
      "step": 18840
    },
    {
      "epoch": 9.876834381551364,
      "grad_norm": 0.6538559794425964,
      "learning_rate": 2.9871650633742198e-05,
      "loss": 0.3864,
      "num_input_tokens_seen": 12316104,
      "step": 18845
    },
    {
      "epoch": 9.879454926624739,
      "grad_norm": 0.7207499146461487,
      "learning_rate": 2.9860435036393142e-05,
      "loss": 0.5438,
      "num_input_tokens_seen": 12318568,
      "step": 18850
    },
    {
      "epoch": 9.882075471698114,
      "grad_norm": 0.6562116146087646,
      "learning_rate": 2.9849218422296527e-05,
      "loss": 0.5038,
      "num_input_tokens_seen": 12321736,
      "step": 18855
    },
    {
      "epoch": 9.884696016771489,
      "grad_norm": 0.4040244221687317,
      "learning_rate": 2.983800079379873e-05,
      "loss": 0.4785,
      "num_input_tokens_seen": 12325032,
      "step": 18860
    },
    {
      "epoch": 9.887316561844864,
      "grad_norm": 0.24109461903572083,
      "learning_rate": 2.982678215324638e-05,
      "loss": 0.4955,
      "num_input_tokens_seen": 12329160,
      "step": 18865
    },
    {
      "epoch": 9.88993710691824,
      "grad_norm": 0.33271074295043945,
      "learning_rate": 2.9815562502986256e-05,
      "loss": 0.6669,
      "num_input_tokens_seen": 12332392,
      "step": 18870
    },
    {
      "epoch": 9.892557651991615,
      "grad_norm": 0.26281359791755676,
      "learning_rate": 2.9804341845365398e-05,
      "loss": 0.4204,
      "num_input_tokens_seen": 12336456,
      "step": 18875
    },
    {
      "epoch": 9.89517819706499,
      "grad_norm": 0.4265877604484558,
      "learning_rate": 2.9793120182731037e-05,
      "loss": 0.4032,
      "num_input_tokens_seen": 12339080,
      "step": 18880
    },
    {
      "epoch": 9.897798742138365,
      "grad_norm": 0.561519205570221,
      "learning_rate": 2.9781897517430617e-05,
      "loss": 0.3848,
      "num_input_tokens_seen": 12341768,
      "step": 18885
    },
    {
      "epoch": 9.90041928721174,
      "grad_norm": 0.526155948638916,
      "learning_rate": 2.977067385181179e-05,
      "loss": 0.4009,
      "num_input_tokens_seen": 12344936,
      "step": 18890
    },
    {
      "epoch": 9.903039832285115,
      "grad_norm": 0.39343002438545227,
      "learning_rate": 2.9759449188222426e-05,
      "loss": 0.4491,
      "num_input_tokens_seen": 12349416,
      "step": 18895
    },
    {
      "epoch": 9.90566037735849,
      "grad_norm": 0.5544989705085754,
      "learning_rate": 2.9748223529010582e-05,
      "loss": 0.493,
      "num_input_tokens_seen": 12352360,
      "step": 18900
    },
    {
      "epoch": 9.908280922431866,
      "grad_norm": 0.4721764028072357,
      "learning_rate": 2.9736996876524546e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 12355688,
      "step": 18905
    },
    {
      "epoch": 9.91090146750524,
      "grad_norm": 0.3213590979576111,
      "learning_rate": 2.9725769233112816e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 12359560,
      "step": 18910
    },
    {
      "epoch": 9.913522012578616,
      "grad_norm": 0.4970027506351471,
      "learning_rate": 2.9714540601124074e-05,
      "loss": 0.5143,
      "num_input_tokens_seen": 12363656,
      "step": 18915
    },
    {
      "epoch": 9.916142557651991,
      "grad_norm": 0.30771130323410034,
      "learning_rate": 2.9703310982907218e-05,
      "loss": 0.3807,
      "num_input_tokens_seen": 12367688,
      "step": 18920
    },
    {
      "epoch": 9.918763102725366,
      "grad_norm": 0.8267582058906555,
      "learning_rate": 2.9692080380811365e-05,
      "loss": 0.4622,
      "num_input_tokens_seen": 12371400,
      "step": 18925
    },
    {
      "epoch": 9.921383647798741,
      "grad_norm": 0.42021018266677856,
      "learning_rate": 2.9680848797185828e-05,
      "loss": 0.436,
      "num_input_tokens_seen": 12374152,
      "step": 18930
    },
    {
      "epoch": 9.924004192872118,
      "grad_norm": 0.6553726196289062,
      "learning_rate": 2.966961623438012e-05,
      "loss": 0.3472,
      "num_input_tokens_seen": 12378472,
      "step": 18935
    },
    {
      "epoch": 9.926624737945493,
      "grad_norm": 0.6036279201507568,
      "learning_rate": 2.9658382694743973e-05,
      "loss": 0.4944,
      "num_input_tokens_seen": 12381480,
      "step": 18940
    },
    {
      "epoch": 9.929245283018869,
      "grad_norm": 0.26560157537460327,
      "learning_rate": 2.9647148180627317e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 12384648,
      "step": 18945
    },
    {
      "epoch": 9.931865828092244,
      "grad_norm": 0.5289164781570435,
      "learning_rate": 2.9635912694380275e-05,
      "loss": 0.3595,
      "num_input_tokens_seen": 12387432,
      "step": 18950
    },
    {
      "epoch": 9.934486373165619,
      "grad_norm": 0.4378933012485504,
      "learning_rate": 2.962467623835319e-05,
      "loss": 0.43,
      "num_input_tokens_seen": 12390984,
      "step": 18955
    },
    {
      "epoch": 9.937106918238994,
      "grad_norm": 0.40249869227409363,
      "learning_rate": 2.961343881489659e-05,
      "loss": 0.3871,
      "num_input_tokens_seen": 12394760,
      "step": 18960
    },
    {
      "epoch": 9.93972746331237,
      "grad_norm": 0.5288019180297852,
      "learning_rate": 2.9602200426361226e-05,
      "loss": 0.5221,
      "num_input_tokens_seen": 12397768,
      "step": 18965
    },
    {
      "epoch": 9.942348008385745,
      "grad_norm": 0.4635269343852997,
      "learning_rate": 2.9590961075098038e-05,
      "loss": 0.3814,
      "num_input_tokens_seen": 12400968,
      "step": 18970
    },
    {
      "epoch": 9.94496855345912,
      "grad_norm": 0.3663213551044464,
      "learning_rate": 2.957972076345817e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 12406664,
      "step": 18975
    },
    {
      "epoch": 9.947589098532495,
      "grad_norm": 0.6126797795295715,
      "learning_rate": 2.9568479493792967e-05,
      "loss": 0.4307,
      "num_input_tokens_seen": 12409448,
      "step": 18980
    },
    {
      "epoch": 9.95020964360587,
      "grad_norm": 0.3024360239505768,
      "learning_rate": 2.955723726845397e-05,
      "loss": 0.3385,
      "num_input_tokens_seen": 12412744,
      "step": 18985
    },
    {
      "epoch": 9.952830188679245,
      "grad_norm": 0.38035038113594055,
      "learning_rate": 2.9545994089792928e-05,
      "loss": 0.6086,
      "num_input_tokens_seen": 12415656,
      "step": 18990
    },
    {
      "epoch": 9.95545073375262,
      "grad_norm": 0.44034361839294434,
      "learning_rate": 2.9534749960161795e-05,
      "loss": 0.5509,
      "num_input_tokens_seen": 12418472,
      "step": 18995
    },
    {
      "epoch": 9.958071278825996,
      "grad_norm": 0.4753412902355194,
      "learning_rate": 2.9523504881912694e-05,
      "loss": 0.396,
      "num_input_tokens_seen": 12421768,
      "step": 19000
    },
    {
      "epoch": 9.96069182389937,
      "grad_norm": 0.41799721121788025,
      "learning_rate": 2.9512258857397978e-05,
      "loss": 0.4354,
      "num_input_tokens_seen": 12425800,
      "step": 19005
    },
    {
      "epoch": 9.963312368972746,
      "grad_norm": 0.7318289279937744,
      "learning_rate": 2.9501011888970203e-05,
      "loss": 0.5472,
      "num_input_tokens_seen": 12429288,
      "step": 19010
    },
    {
      "epoch": 9.965932914046121,
      "grad_norm": 0.4493831396102905,
      "learning_rate": 2.9489763978982075e-05,
      "loss": 0.3823,
      "num_input_tokens_seen": 12433160,
      "step": 19015
    },
    {
      "epoch": 9.968553459119496,
      "grad_norm": 1.2685402631759644,
      "learning_rate": 2.9478515129786556e-05,
      "loss": 0.4388,
      "num_input_tokens_seen": 12435944,
      "step": 19020
    },
    {
      "epoch": 9.971174004192871,
      "grad_norm": 0.3756071925163269,
      "learning_rate": 2.946726534373676e-05,
      "loss": 0.3875,
      "num_input_tokens_seen": 12438216,
      "step": 19025
    },
    {
      "epoch": 9.973794549266248,
      "grad_norm": 0.6779617071151733,
      "learning_rate": 2.945601462318601e-05,
      "loss": 0.4013,
      "num_input_tokens_seen": 12440904,
      "step": 19030
    },
    {
      "epoch": 9.976415094339622,
      "grad_norm": 0.7233598828315735,
      "learning_rate": 2.944476297048786e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 12443752,
      "step": 19035
    },
    {
      "epoch": 9.979035639412999,
      "grad_norm": 0.9422183036804199,
      "learning_rate": 2.9433510387995987e-05,
      "loss": 0.6508,
      "num_input_tokens_seen": 12446152,
      "step": 19040
    },
    {
      "epoch": 9.981656184486374,
      "grad_norm": 0.4371408522129059,
      "learning_rate": 2.9422256878064325e-05,
      "loss": 0.4446,
      "num_input_tokens_seen": 12449160,
      "step": 19045
    },
    {
      "epoch": 9.984276729559749,
      "grad_norm": 0.40764105319976807,
      "learning_rate": 2.9411002443046982e-05,
      "loss": 0.4862,
      "num_input_tokens_seen": 12453064,
      "step": 19050
    },
    {
      "epoch": 9.986897274633124,
      "grad_norm": 0.7701849937438965,
      "learning_rate": 2.9399747085298245e-05,
      "loss": 0.4392,
      "num_input_tokens_seen": 12455944,
      "step": 19055
    },
    {
      "epoch": 9.9895178197065,
      "grad_norm": 0.5449591875076294,
      "learning_rate": 2.9388490807172614e-05,
      "loss": 0.4937,
      "num_input_tokens_seen": 12459528,
      "step": 19060
    },
    {
      "epoch": 9.992138364779874,
      "grad_norm": 0.3166823983192444,
      "learning_rate": 2.9377233611024775e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 12463048,
      "step": 19065
    },
    {
      "epoch": 9.99475890985325,
      "grad_norm": 0.48080092668533325,
      "learning_rate": 2.9365975499209596e-05,
      "loss": 0.4336,
      "num_input_tokens_seen": 12467240,
      "step": 19070
    },
    {
      "epoch": 9.997379454926625,
      "grad_norm": 0.4152337312698364,
      "learning_rate": 2.9354716474082154e-05,
      "loss": 0.6285,
      "num_input_tokens_seen": 12470216,
      "step": 19075
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.3445322513580322,
      "learning_rate": 2.9343456537997714e-05,
      "loss": 0.6523,
      "num_input_tokens_seen": 12472912,
      "step": 19080
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.4766788184642792,
      "eval_runtime": 9.8336,
      "eval_samples_per_second": 86.235,
      "eval_steps_per_second": 21.559,
      "num_input_tokens_seen": 12472912,
      "step": 19080
    },
    {
      "epoch": 10.002620545073375,
      "grad_norm": 0.569311261177063,
      "learning_rate": 2.9332195693311714e-05,
      "loss": 0.3735,
      "num_input_tokens_seen": 12475568,
      "step": 19085
    },
    {
      "epoch": 10.00524109014675,
      "grad_norm": 1.5170177221298218,
      "learning_rate": 2.9320933942379803e-05,
      "loss": 0.4064,
      "num_input_tokens_seen": 12478128,
      "step": 19090
    },
    {
      "epoch": 10.007861635220126,
      "grad_norm": 0.5511986017227173,
      "learning_rate": 2.9309671287557805e-05,
      "loss": 0.4138,
      "num_input_tokens_seen": 12481520,
      "step": 19095
    },
    {
      "epoch": 10.0104821802935,
      "grad_norm": 0.2923735976219177,
      "learning_rate": 2.9298407731201737e-05,
      "loss": 0.5298,
      "num_input_tokens_seen": 12485264,
      "step": 19100
    },
    {
      "epoch": 10.013102725366876,
      "grad_norm": 0.4232950508594513,
      "learning_rate": 2.928714327566781e-05,
      "loss": 0.3755,
      "num_input_tokens_seen": 12488464,
      "step": 19105
    },
    {
      "epoch": 10.015723270440251,
      "grad_norm": 1.1852926015853882,
      "learning_rate": 2.9275877923312427e-05,
      "loss": 0.4191,
      "num_input_tokens_seen": 12491504,
      "step": 19110
    },
    {
      "epoch": 10.018343815513626,
      "grad_norm": 0.7056534290313721,
      "learning_rate": 2.9264611676492165e-05,
      "loss": 0.7435,
      "num_input_tokens_seen": 12494576,
      "step": 19115
    },
    {
      "epoch": 10.020964360587001,
      "grad_norm": 0.4768909811973572,
      "learning_rate": 2.9253344537563782e-05,
      "loss": 0.4648,
      "num_input_tokens_seen": 12497648,
      "step": 19120
    },
    {
      "epoch": 10.023584905660377,
      "grad_norm": 0.5906173586845398,
      "learning_rate": 2.924207650888426e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 12500208,
      "step": 19125
    },
    {
      "epoch": 10.026205450733753,
      "grad_norm": 0.5999085307121277,
      "learning_rate": 2.9230807592810723e-05,
      "loss": 0.5453,
      "num_input_tokens_seen": 12503248,
      "step": 19130
    },
    {
      "epoch": 10.028825995807129,
      "grad_norm": 0.2785259783267975,
      "learning_rate": 2.92195377917005e-05,
      "loss": 0.4394,
      "num_input_tokens_seen": 12508080,
      "step": 19135
    },
    {
      "epoch": 10.031446540880504,
      "grad_norm": 0.3026297986507416,
      "learning_rate": 2.920826710791112e-05,
      "loss": 0.4084,
      "num_input_tokens_seen": 12511856,
      "step": 19140
    },
    {
      "epoch": 10.034067085953879,
      "grad_norm": 0.36950066685676575,
      "learning_rate": 2.919699554380026e-05,
      "loss": 0.4142,
      "num_input_tokens_seen": 12515920,
      "step": 19145
    },
    {
      "epoch": 10.036687631027254,
      "grad_norm": 0.575124204158783,
      "learning_rate": 2.9185723101725825e-05,
      "loss": 0.5099,
      "num_input_tokens_seen": 12519024,
      "step": 19150
    },
    {
      "epoch": 10.03930817610063,
      "grad_norm": 0.588796079158783,
      "learning_rate": 2.9174449784045867e-05,
      "loss": 0.4981,
      "num_input_tokens_seen": 12521840,
      "step": 19155
    },
    {
      "epoch": 10.041928721174004,
      "grad_norm": 0.46534109115600586,
      "learning_rate": 2.916317559311863e-05,
      "loss": 0.4926,
      "num_input_tokens_seen": 12525008,
      "step": 19160
    },
    {
      "epoch": 10.04454926624738,
      "grad_norm": 0.34193259477615356,
      "learning_rate": 2.9151900531302566e-05,
      "loss": 0.3891,
      "num_input_tokens_seen": 12528048,
      "step": 19165
    },
    {
      "epoch": 10.047169811320755,
      "grad_norm": 0.5669025182723999,
      "learning_rate": 2.9140624600956272e-05,
      "loss": 0.3964,
      "num_input_tokens_seen": 12531536,
      "step": 19170
    },
    {
      "epoch": 10.04979035639413,
      "grad_norm": 0.43315389752388,
      "learning_rate": 2.912934780443855e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 12535248,
      "step": 19175
    },
    {
      "epoch": 10.052410901467505,
      "grad_norm": 0.37110430002212524,
      "learning_rate": 2.9118070144108388e-05,
      "loss": 0.3652,
      "num_input_tokens_seen": 12538128,
      "step": 19180
    },
    {
      "epoch": 10.05503144654088,
      "grad_norm": 0.39148637652397156,
      "learning_rate": 2.9106791622324924e-05,
      "loss": 0.403,
      "num_input_tokens_seen": 12541136,
      "step": 19185
    },
    {
      "epoch": 10.057651991614255,
      "grad_norm": 0.3167612850666046,
      "learning_rate": 2.909551224144751e-05,
      "loss": 0.4219,
      "num_input_tokens_seen": 12544272,
      "step": 19190
    },
    {
      "epoch": 10.06027253668763,
      "grad_norm": 0.47915196418762207,
      "learning_rate": 2.9084232003835653e-05,
      "loss": 0.4183,
      "num_input_tokens_seen": 12547184,
      "step": 19195
    },
    {
      "epoch": 10.062893081761006,
      "grad_norm": 0.3839564323425293,
      "learning_rate": 2.9072950911849058e-05,
      "loss": 0.4238,
      "num_input_tokens_seen": 12550416,
      "step": 19200
    },
    {
      "epoch": 10.065513626834381,
      "grad_norm": 0.4822752773761749,
      "learning_rate": 2.90616689678476e-05,
      "loss": 0.4154,
      "num_input_tokens_seen": 12552688,
      "step": 19205
    },
    {
      "epoch": 10.068134171907756,
      "grad_norm": 0.4564165771007538,
      "learning_rate": 2.9050386174191337e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 12555536,
      "step": 19210
    },
    {
      "epoch": 10.070754716981131,
      "grad_norm": 0.38704633712768555,
      "learning_rate": 2.903910253324048e-05,
      "loss": 0.4719,
      "num_input_tokens_seen": 12558672,
      "step": 19215
    },
    {
      "epoch": 10.073375262054507,
      "grad_norm": 0.7942884564399719,
      "learning_rate": 2.9027818047355477e-05,
      "loss": 0.4494,
      "num_input_tokens_seen": 12562192,
      "step": 19220
    },
    {
      "epoch": 10.075995807127883,
      "grad_norm": 0.433870404958725,
      "learning_rate": 2.9016532718896873e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 12565584,
      "step": 19225
    },
    {
      "epoch": 10.078616352201259,
      "grad_norm": 0.4810861349105835,
      "learning_rate": 2.900524655022545e-05,
      "loss": 0.4394,
      "num_input_tokens_seen": 12568720,
      "step": 19230
    },
    {
      "epoch": 10.081236897274634,
      "grad_norm": 0.4349820911884308,
      "learning_rate": 2.899395954370215e-05,
      "loss": 0.3564,
      "num_input_tokens_seen": 12572432,
      "step": 19235
    },
    {
      "epoch": 10.083857442348009,
      "grad_norm": 0.5374131202697754,
      "learning_rate": 2.898267170168807e-05,
      "loss": 0.4714,
      "num_input_tokens_seen": 12575632,
      "step": 19240
    },
    {
      "epoch": 10.086477987421384,
      "grad_norm": 0.26599499583244324,
      "learning_rate": 2.8971383026544507e-05,
      "loss": 0.4208,
      "num_input_tokens_seen": 12579888,
      "step": 19245
    },
    {
      "epoch": 10.08909853249476,
      "grad_norm": 0.6897318363189697,
      "learning_rate": 2.896009352063293e-05,
      "loss": 0.3641,
      "num_input_tokens_seen": 12582416,
      "step": 19250
    },
    {
      "epoch": 10.091719077568134,
      "grad_norm": 0.3512917160987854,
      "learning_rate": 2.894880318631496e-05,
      "loss": 0.3825,
      "num_input_tokens_seen": 12585360,
      "step": 19255
    },
    {
      "epoch": 10.09433962264151,
      "grad_norm": 0.2757076621055603,
      "learning_rate": 2.8937512025952417e-05,
      "loss": 0.4191,
      "num_input_tokens_seen": 12589456,
      "step": 19260
    },
    {
      "epoch": 10.096960167714885,
      "grad_norm": 0.3709339499473572,
      "learning_rate": 2.8926220041907283e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 12592752,
      "step": 19265
    },
    {
      "epoch": 10.09958071278826,
      "grad_norm": 0.38438838720321655,
      "learning_rate": 2.8914927236541706e-05,
      "loss": 0.5482,
      "num_input_tokens_seen": 12596304,
      "step": 19270
    },
    {
      "epoch": 10.102201257861635,
      "grad_norm": 0.5277635455131531,
      "learning_rate": 2.8903633612218012e-05,
      "loss": 0.4893,
      "num_input_tokens_seen": 12600048,
      "step": 19275
    },
    {
      "epoch": 10.10482180293501,
      "grad_norm": 0.4197789430618286,
      "learning_rate": 2.8892339171298703e-05,
      "loss": 0.3735,
      "num_input_tokens_seen": 12602992,
      "step": 19280
    },
    {
      "epoch": 10.107442348008385,
      "grad_norm": 0.32678139209747314,
      "learning_rate": 2.888104391614645e-05,
      "loss": 0.4456,
      "num_input_tokens_seen": 12606960,
      "step": 19285
    },
    {
      "epoch": 10.11006289308176,
      "grad_norm": 0.4232761561870575,
      "learning_rate": 2.886974784912408e-05,
      "loss": 0.3503,
      "num_input_tokens_seen": 12609584,
      "step": 19290
    },
    {
      "epoch": 10.112683438155136,
      "grad_norm": 0.3381011486053467,
      "learning_rate": 2.8858450972594612e-05,
      "loss": 0.4513,
      "num_input_tokens_seen": 12612368,
      "step": 19295
    },
    {
      "epoch": 10.115303983228511,
      "grad_norm": 0.5074424147605896,
      "learning_rate": 2.8847153288921226e-05,
      "loss": 0.4072,
      "num_input_tokens_seen": 12615472,
      "step": 19300
    },
    {
      "epoch": 10.117924528301886,
      "grad_norm": 0.44946637749671936,
      "learning_rate": 2.883585480046725e-05,
      "loss": 0.6293,
      "num_input_tokens_seen": 12621456,
      "step": 19305
    },
    {
      "epoch": 10.120545073375261,
      "grad_norm": 0.7455275654792786,
      "learning_rate": 2.8824555509596223e-05,
      "loss": 0.4232,
      "num_input_tokens_seen": 12624624,
      "step": 19310
    },
    {
      "epoch": 10.123165618448636,
      "grad_norm": 0.4230788052082062,
      "learning_rate": 2.8813255418671815e-05,
      "loss": 0.5311,
      "num_input_tokens_seen": 12628112,
      "step": 19315
    },
    {
      "epoch": 10.125786163522013,
      "grad_norm": 0.41270822286605835,
      "learning_rate": 2.8801954530057874e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 12631536,
      "step": 19320
    },
    {
      "epoch": 10.128406708595389,
      "grad_norm": 0.8229631781578064,
      "learning_rate": 2.879065284611842e-05,
      "loss": 0.4511,
      "num_input_tokens_seen": 12634704,
      "step": 19325
    },
    {
      "epoch": 10.131027253668764,
      "grad_norm": 0.3454727530479431,
      "learning_rate": 2.8779350369217634e-05,
      "loss": 0.4219,
      "num_input_tokens_seen": 12638640,
      "step": 19330
    },
    {
      "epoch": 10.133647798742139,
      "grad_norm": 0.3981066942214966,
      "learning_rate": 2.876804710171987e-05,
      "loss": 0.3913,
      "num_input_tokens_seen": 12641392,
      "step": 19335
    },
    {
      "epoch": 10.136268343815514,
      "grad_norm": 0.2818308472633362,
      "learning_rate": 2.8756743045989643e-05,
      "loss": 0.3387,
      "num_input_tokens_seen": 12648656,
      "step": 19340
    },
    {
      "epoch": 10.13888888888889,
      "grad_norm": 0.6821882724761963,
      "learning_rate": 2.8745438204391623e-05,
      "loss": 0.3804,
      "num_input_tokens_seen": 12651760,
      "step": 19345
    },
    {
      "epoch": 10.141509433962264,
      "grad_norm": 0.7245747447013855,
      "learning_rate": 2.8734132579290675e-05,
      "loss": 0.4608,
      "num_input_tokens_seen": 12655120,
      "step": 19350
    },
    {
      "epoch": 10.14412997903564,
      "grad_norm": 1.0757780075073242,
      "learning_rate": 2.8722826173051782e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 12657936,
      "step": 19355
    },
    {
      "epoch": 10.146750524109015,
      "grad_norm": 0.6926081776618958,
      "learning_rate": 2.871151898804012e-05,
      "loss": 0.4131,
      "num_input_tokens_seen": 12661648,
      "step": 19360
    },
    {
      "epoch": 10.14937106918239,
      "grad_norm": 0.42891940474510193,
      "learning_rate": 2.8700211026621043e-05,
      "loss": 0.6113,
      "num_input_tokens_seen": 12665328,
      "step": 19365
    },
    {
      "epoch": 10.151991614255765,
      "grad_norm": 0.5518250465393066,
      "learning_rate": 2.8688902291160025e-05,
      "loss": 0.4361,
      "num_input_tokens_seen": 12668464,
      "step": 19370
    },
    {
      "epoch": 10.15461215932914,
      "grad_norm": 0.4421013295650482,
      "learning_rate": 2.867759278402273e-05,
      "loss": 0.3772,
      "num_input_tokens_seen": 12672016,
      "step": 19375
    },
    {
      "epoch": 10.157232704402515,
      "grad_norm": 0.7472148537635803,
      "learning_rate": 2.8666282507574994e-05,
      "loss": 0.503,
      "num_input_tokens_seen": 12675312,
      "step": 19380
    },
    {
      "epoch": 10.15985324947589,
      "grad_norm": 0.6283725500106812,
      "learning_rate": 2.8654971464182773e-05,
      "loss": 0.6054,
      "num_input_tokens_seen": 12677872,
      "step": 19385
    },
    {
      "epoch": 10.162473794549266,
      "grad_norm": 0.5284227132797241,
      "learning_rate": 2.8643659656212222e-05,
      "loss": 0.4014,
      "num_input_tokens_seen": 12684112,
      "step": 19390
    },
    {
      "epoch": 10.165094339622641,
      "grad_norm": 0.4495986998081207,
      "learning_rate": 2.8632347086029642e-05,
      "loss": 0.3004,
      "num_input_tokens_seen": 12686736,
      "step": 19395
    },
    {
      "epoch": 10.167714884696016,
      "grad_norm": 0.2231825441122055,
      "learning_rate": 2.8621033756001486e-05,
      "loss": 0.6308,
      "num_input_tokens_seen": 12690352,
      "step": 19400
    },
    {
      "epoch": 10.170335429769391,
      "grad_norm": 0.38539016246795654,
      "learning_rate": 2.8609719668494394e-05,
      "loss": 0.4191,
      "num_input_tokens_seen": 12693264,
      "step": 19405
    },
    {
      "epoch": 10.172955974842766,
      "grad_norm": 0.36369410157203674,
      "learning_rate": 2.8598404825875113e-05,
      "loss": 0.3812,
      "num_input_tokens_seen": 12695888,
      "step": 19410
    },
    {
      "epoch": 10.175576519916143,
      "grad_norm": 0.503027081489563,
      "learning_rate": 2.8587089230510604e-05,
      "loss": 0.4876,
      "num_input_tokens_seen": 12699216,
      "step": 19415
    },
    {
      "epoch": 10.178197064989519,
      "grad_norm": 0.3665699362754822,
      "learning_rate": 2.857577288476795e-05,
      "loss": 0.3966,
      "num_input_tokens_seen": 12703344,
      "step": 19420
    },
    {
      "epoch": 10.180817610062894,
      "grad_norm": 0.3501930832862854,
      "learning_rate": 2.85644557910144e-05,
      "loss": 0.3785,
      "num_input_tokens_seen": 12706480,
      "step": 19425
    },
    {
      "epoch": 10.183438155136269,
      "grad_norm": 0.46877244114875793,
      "learning_rate": 2.8553137951617366e-05,
      "loss": 0.4551,
      "num_input_tokens_seen": 12709744,
      "step": 19430
    },
    {
      "epoch": 10.186058700209644,
      "grad_norm": 0.35728251934051514,
      "learning_rate": 2.854181936894441e-05,
      "loss": 0.511,
      "num_input_tokens_seen": 12712816,
      "step": 19435
    },
    {
      "epoch": 10.18867924528302,
      "grad_norm": 0.49717098474502563,
      "learning_rate": 2.8530500045363247e-05,
      "loss": 0.4583,
      "num_input_tokens_seen": 12716304,
      "step": 19440
    },
    {
      "epoch": 10.191299790356394,
      "grad_norm": 0.3807353973388672,
      "learning_rate": 2.851917998324175e-05,
      "loss": 0.4148,
      "num_input_tokens_seen": 12719216,
      "step": 19445
    },
    {
      "epoch": 10.19392033542977,
      "grad_norm": 0.4867483973503113,
      "learning_rate": 2.8507859184947957e-05,
      "loss": 0.5286,
      "num_input_tokens_seen": 12723312,
      "step": 19450
    },
    {
      "epoch": 10.196540880503145,
      "grad_norm": 0.2578546404838562,
      "learning_rate": 2.8496537652850043e-05,
      "loss": 0.4057,
      "num_input_tokens_seen": 12726320,
      "step": 19455
    },
    {
      "epoch": 10.19916142557652,
      "grad_norm": 0.5190029740333557,
      "learning_rate": 2.8485215389316333e-05,
      "loss": 0.549,
      "num_input_tokens_seen": 12730128,
      "step": 19460
    },
    {
      "epoch": 10.201781970649895,
      "grad_norm": 0.2739211916923523,
      "learning_rate": 2.8473892396715328e-05,
      "loss": 0.6547,
      "num_input_tokens_seen": 12735024,
      "step": 19465
    },
    {
      "epoch": 10.20440251572327,
      "grad_norm": 0.36935874819755554,
      "learning_rate": 2.846256867741567e-05,
      "loss": 0.3094,
      "num_input_tokens_seen": 12737872,
      "step": 19470
    },
    {
      "epoch": 10.207023060796645,
      "grad_norm": 0.4665168523788452,
      "learning_rate": 2.845124423378614e-05,
      "loss": 0.4061,
      "num_input_tokens_seen": 12740400,
      "step": 19475
    },
    {
      "epoch": 10.20964360587002,
      "grad_norm": 0.28829479217529297,
      "learning_rate": 2.8439919068195697e-05,
      "loss": 0.3175,
      "num_input_tokens_seen": 12743664,
      "step": 19480
    },
    {
      "epoch": 10.212264150943396,
      "grad_norm": 0.4745318591594696,
      "learning_rate": 2.842859318301343e-05,
      "loss": 0.5755,
      "num_input_tokens_seen": 12746512,
      "step": 19485
    },
    {
      "epoch": 10.214884696016771,
      "grad_norm": 0.2901303768157959,
      "learning_rate": 2.841726658060858e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 12750160,
      "step": 19490
    },
    {
      "epoch": 10.217505241090146,
      "grad_norm": 0.2652429938316345,
      "learning_rate": 2.8405939263350555e-05,
      "loss": 0.44,
      "num_input_tokens_seen": 12753520,
      "step": 19495
    },
    {
      "epoch": 10.220125786163521,
      "grad_norm": 0.2906486392021179,
      "learning_rate": 2.8394611233608886e-05,
      "loss": 0.4583,
      "num_input_tokens_seen": 12756624,
      "step": 19500
    },
    {
      "epoch": 10.222746331236896,
      "grad_norm": 0.5216783881187439,
      "learning_rate": 2.8383282493753283e-05,
      "loss": 0.4052,
      "num_input_tokens_seen": 12759248,
      "step": 19505
    },
    {
      "epoch": 10.225366876310272,
      "grad_norm": 0.43340668082237244,
      "learning_rate": 2.8371953046153575e-05,
      "loss": 0.3629,
      "num_input_tokens_seen": 12762384,
      "step": 19510
    },
    {
      "epoch": 10.227987421383649,
      "grad_norm": 0.4376310110092163,
      "learning_rate": 2.8360622893179757e-05,
      "loss": 0.4959,
      "num_input_tokens_seen": 12766224,
      "step": 19515
    },
    {
      "epoch": 10.230607966457024,
      "grad_norm": 0.4058534801006317,
      "learning_rate": 2.834929203720197e-05,
      "loss": 0.4317,
      "num_input_tokens_seen": 12768912,
      "step": 19520
    },
    {
      "epoch": 10.233228511530399,
      "grad_norm": 0.36926716566085815,
      "learning_rate": 2.83379604805905e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 12772336,
      "step": 19525
    },
    {
      "epoch": 10.235849056603774,
      "grad_norm": 0.8545995950698853,
      "learning_rate": 2.832662822571578e-05,
      "loss": 0.3847,
      "num_input_tokens_seen": 12775216,
      "step": 19530
    },
    {
      "epoch": 10.23846960167715,
      "grad_norm": 0.421647846698761,
      "learning_rate": 2.8315295274948395e-05,
      "loss": 0.4173,
      "num_input_tokens_seen": 12778512,
      "step": 19535
    },
    {
      "epoch": 10.241090146750524,
      "grad_norm": 0.5413495302200317,
      "learning_rate": 2.8303961630659047e-05,
      "loss": 0.3415,
      "num_input_tokens_seen": 12782576,
      "step": 19540
    },
    {
      "epoch": 10.2437106918239,
      "grad_norm": 0.41547060012817383,
      "learning_rate": 2.8292627295218627e-05,
      "loss": 0.4036,
      "num_input_tokens_seen": 12785584,
      "step": 19545
    },
    {
      "epoch": 10.246331236897275,
      "grad_norm": 0.37010860443115234,
      "learning_rate": 2.828129227099815e-05,
      "loss": 0.4252,
      "num_input_tokens_seen": 12788432,
      "step": 19550
    },
    {
      "epoch": 10.24895178197065,
      "grad_norm": 0.717210590839386,
      "learning_rate": 2.826995656036875e-05,
      "loss": 0.6719,
      "num_input_tokens_seen": 12791312,
      "step": 19555
    },
    {
      "epoch": 10.251572327044025,
      "grad_norm": 0.328844279050827,
      "learning_rate": 2.8258620165701747e-05,
      "loss": 0.384,
      "num_input_tokens_seen": 12795120,
      "step": 19560
    },
    {
      "epoch": 10.2541928721174,
      "grad_norm": 0.5662649273872375,
      "learning_rate": 2.8247283089368586e-05,
      "loss": 0.3775,
      "num_input_tokens_seen": 12798032,
      "step": 19565
    },
    {
      "epoch": 10.256813417190775,
      "grad_norm": 0.6349830031394958,
      "learning_rate": 2.8235945333740842e-05,
      "loss": 0.4198,
      "num_input_tokens_seen": 12800688,
      "step": 19570
    },
    {
      "epoch": 10.25943396226415,
      "grad_norm": 0.27608928084373474,
      "learning_rate": 2.822460690119026e-05,
      "loss": 0.5137,
      "num_input_tokens_seen": 12804240,
      "step": 19575
    },
    {
      "epoch": 10.262054507337526,
      "grad_norm": 0.5525272488594055,
      "learning_rate": 2.82132677940887e-05,
      "loss": 0.4543,
      "num_input_tokens_seen": 12807216,
      "step": 19580
    },
    {
      "epoch": 10.264675052410901,
      "grad_norm": 0.4366496205329895,
      "learning_rate": 2.820192801480817e-05,
      "loss": 0.3988,
      "num_input_tokens_seen": 12810352,
      "step": 19585
    },
    {
      "epoch": 10.267295597484276,
      "grad_norm": 0.5281155705451965,
      "learning_rate": 2.8190587565720842e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 12813328,
      "step": 19590
    },
    {
      "epoch": 10.269916142557651,
      "grad_norm": 0.4916626214981079,
      "learning_rate": 2.8179246449198986e-05,
      "loss": 0.4508,
      "num_input_tokens_seen": 12816336,
      "step": 19595
    },
    {
      "epoch": 10.272536687631026,
      "grad_norm": 0.7577579021453857,
      "learning_rate": 2.8167904667615048e-05,
      "loss": 0.3883,
      "num_input_tokens_seen": 12820624,
      "step": 19600
    },
    {
      "epoch": 10.275157232704402,
      "grad_norm": 0.38692352175712585,
      "learning_rate": 2.8156562223341592e-05,
      "loss": 0.5015,
      "num_input_tokens_seen": 12824176,
      "step": 19605
    },
    {
      "epoch": 10.277777777777779,
      "grad_norm": 0.6406793594360352,
      "learning_rate": 2.814521911875133e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 12827376,
      "step": 19610
    },
    {
      "epoch": 10.280398322851154,
      "grad_norm": 0.5908503532409668,
      "learning_rate": 2.8133875356217114e-05,
      "loss": 0.3022,
      "num_input_tokens_seen": 12830704,
      "step": 19615
    },
    {
      "epoch": 10.283018867924529,
      "grad_norm": 0.5460635423660278,
      "learning_rate": 2.8122530938111923e-05,
      "loss": 0.5455,
      "num_input_tokens_seen": 12833616,
      "step": 19620
    },
    {
      "epoch": 10.285639412997904,
      "grad_norm": 0.33490365743637085,
      "learning_rate": 2.8111185866808888e-05,
      "loss": 0.4647,
      "num_input_tokens_seen": 12836688,
      "step": 19625
    },
    {
      "epoch": 10.28825995807128,
      "grad_norm": 0.36455556750297546,
      "learning_rate": 2.809984014468126e-05,
      "loss": 0.4964,
      "num_input_tokens_seen": 12840656,
      "step": 19630
    },
    {
      "epoch": 10.290880503144654,
      "grad_norm": 0.3549049496650696,
      "learning_rate": 2.8088493774102443e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 12843408,
      "step": 19635
    },
    {
      "epoch": 10.29350104821803,
      "grad_norm": 0.5526121854782104,
      "learning_rate": 2.807714675744596e-05,
      "loss": 0.5389,
      "num_input_tokens_seen": 12846224,
      "step": 19640
    },
    {
      "epoch": 10.296121593291405,
      "grad_norm": 0.6429850459098816,
      "learning_rate": 2.8065799097085475e-05,
      "loss": 0.5309,
      "num_input_tokens_seen": 12850288,
      "step": 19645
    },
    {
      "epoch": 10.29874213836478,
      "grad_norm": 0.5608636736869812,
      "learning_rate": 2.8054450795394804e-05,
      "loss": 0.4625,
      "num_input_tokens_seen": 12853424,
      "step": 19650
    },
    {
      "epoch": 10.301362683438155,
      "grad_norm": 0.7649051547050476,
      "learning_rate": 2.8043101854747878e-05,
      "loss": 0.4894,
      "num_input_tokens_seen": 12856656,
      "step": 19655
    },
    {
      "epoch": 10.30398322851153,
      "grad_norm": 0.3625293970108032,
      "learning_rate": 2.8031752277518748e-05,
      "loss": 0.5196,
      "num_input_tokens_seen": 12859248,
      "step": 19660
    },
    {
      "epoch": 10.306603773584905,
      "grad_norm": 0.44268447160720825,
      "learning_rate": 2.802040206608164e-05,
      "loss": 0.3594,
      "num_input_tokens_seen": 12862160,
      "step": 19665
    },
    {
      "epoch": 10.30922431865828,
      "grad_norm": 0.9327974915504456,
      "learning_rate": 2.8009051222810877e-05,
      "loss": 0.5213,
      "num_input_tokens_seen": 12864304,
      "step": 19670
    },
    {
      "epoch": 10.311844863731656,
      "grad_norm": 0.3819923996925354,
      "learning_rate": 2.7997699750080918e-05,
      "loss": 0.4971,
      "num_input_tokens_seen": 12868752,
      "step": 19675
    },
    {
      "epoch": 10.314465408805031,
      "grad_norm": 0.42659205198287964,
      "learning_rate": 2.7986347650266386e-05,
      "loss": 0.434,
      "num_input_tokens_seen": 12872112,
      "step": 19680
    },
    {
      "epoch": 10.317085953878406,
      "grad_norm": 0.6323944926261902,
      "learning_rate": 2.797499492574199e-05,
      "loss": 0.4268,
      "num_input_tokens_seen": 12874576,
      "step": 19685
    },
    {
      "epoch": 10.319706498951781,
      "grad_norm": 0.5095192790031433,
      "learning_rate": 2.7963641578882594e-05,
      "loss": 0.4566,
      "num_input_tokens_seen": 12877232,
      "step": 19690
    },
    {
      "epoch": 10.322327044025156,
      "grad_norm": 0.6817609071731567,
      "learning_rate": 2.7952287612063193e-05,
      "loss": 0.406,
      "num_input_tokens_seen": 12880656,
      "step": 19695
    },
    {
      "epoch": 10.324947589098532,
      "grad_norm": 0.4512677788734436,
      "learning_rate": 2.79409330276589e-05,
      "loss": 0.3775,
      "num_input_tokens_seen": 12883472,
      "step": 19700
    },
    {
      "epoch": 10.327568134171909,
      "grad_norm": 0.568718671798706,
      "learning_rate": 2.7929577828044974e-05,
      "loss": 0.4185,
      "num_input_tokens_seen": 12885776,
      "step": 19705
    },
    {
      "epoch": 10.330188679245284,
      "grad_norm": 0.6502196192741394,
      "learning_rate": 2.7918222015596784e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 12888688,
      "step": 19710
    },
    {
      "epoch": 10.332809224318659,
      "grad_norm": 0.6531786322593689,
      "learning_rate": 2.7906865592689847e-05,
      "loss": 0.4925,
      "num_input_tokens_seen": 12891664,
      "step": 19715
    },
    {
      "epoch": 10.335429769392034,
      "grad_norm": 0.8783172965049744,
      "learning_rate": 2.789550856169979e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 12894832,
      "step": 19720
    },
    {
      "epoch": 10.33805031446541,
      "grad_norm": 0.30180326104164124,
      "learning_rate": 2.7884150925002374e-05,
      "loss": 0.316,
      "num_input_tokens_seen": 12897744,
      "step": 19725
    },
    {
      "epoch": 10.340670859538784,
      "grad_norm": 0.5943984389305115,
      "learning_rate": 2.787279268497349e-05,
      "loss": 0.4114,
      "num_input_tokens_seen": 12901008,
      "step": 19730
    },
    {
      "epoch": 10.34329140461216,
      "grad_norm": 0.3661918342113495,
      "learning_rate": 2.786143384398915e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 12903792,
      "step": 19735
    },
    {
      "epoch": 10.345911949685535,
      "grad_norm": 0.5250091552734375,
      "learning_rate": 2.785007440442549e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 12907120,
      "step": 19740
    },
    {
      "epoch": 10.34853249475891,
      "grad_norm": 0.7811464071273804,
      "learning_rate": 2.7838714368658785e-05,
      "loss": 0.5352,
      "num_input_tokens_seen": 12910320,
      "step": 19745
    },
    {
      "epoch": 10.351153039832285,
      "grad_norm": 0.5756775140762329,
      "learning_rate": 2.7827353739065427e-05,
      "loss": 0.4892,
      "num_input_tokens_seen": 12913008,
      "step": 19750
    },
    {
      "epoch": 10.35377358490566,
      "grad_norm": 0.44417667388916016,
      "learning_rate": 2.781599251802191e-05,
      "loss": 0.5375,
      "num_input_tokens_seen": 12916432,
      "step": 19755
    },
    {
      "epoch": 10.356394129979035,
      "grad_norm": 0.4642179608345032,
      "learning_rate": 2.780463070790489e-05,
      "loss": 0.4578,
      "num_input_tokens_seen": 12919536,
      "step": 19760
    },
    {
      "epoch": 10.35901467505241,
      "grad_norm": 0.29900237917900085,
      "learning_rate": 2.7793268311091125e-05,
      "loss": 0.4113,
      "num_input_tokens_seen": 12922768,
      "step": 19765
    },
    {
      "epoch": 10.361635220125786,
      "grad_norm": 0.46073898673057556,
      "learning_rate": 2.7781905329957496e-05,
      "loss": 0.4195,
      "num_input_tokens_seen": 12925840,
      "step": 19770
    },
    {
      "epoch": 10.364255765199161,
      "grad_norm": 0.5667046308517456,
      "learning_rate": 2.7770541766881008e-05,
      "loss": 0.3777,
      "num_input_tokens_seen": 12928656,
      "step": 19775
    },
    {
      "epoch": 10.366876310272536,
      "grad_norm": 0.4206542670726776,
      "learning_rate": 2.7759177624238786e-05,
      "loss": 0.4465,
      "num_input_tokens_seen": 12931504,
      "step": 19780
    },
    {
      "epoch": 10.369496855345911,
      "grad_norm": 0.33535075187683105,
      "learning_rate": 2.7747812904408094e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 12935056,
      "step": 19785
    },
    {
      "epoch": 10.372117400419286,
      "grad_norm": 0.885070264339447,
      "learning_rate": 2.7736447609766286e-05,
      "loss": 0.4054,
      "num_input_tokens_seen": 12938064,
      "step": 19790
    },
    {
      "epoch": 10.374737945492662,
      "grad_norm": 0.3799631595611572,
      "learning_rate": 2.772508174269085e-05,
      "loss": 0.466,
      "num_input_tokens_seen": 12941456,
      "step": 19795
    },
    {
      "epoch": 10.377358490566039,
      "grad_norm": 0.4979081451892853,
      "learning_rate": 2.7713715305559414e-05,
      "loss": 0.4066,
      "num_input_tokens_seen": 12945232,
      "step": 19800
    },
    {
      "epoch": 10.379979035639414,
      "grad_norm": 0.499384343624115,
      "learning_rate": 2.7702348300749692e-05,
      "loss": 0.3558,
      "num_input_tokens_seen": 12947760,
      "step": 19805
    },
    {
      "epoch": 10.382599580712789,
      "grad_norm": 0.3375655710697174,
      "learning_rate": 2.7690980730639533e-05,
      "loss": 0.4173,
      "num_input_tokens_seen": 12951216,
      "step": 19810
    },
    {
      "epoch": 10.385220125786164,
      "grad_norm": 0.4047258198261261,
      "learning_rate": 2.7679612597606902e-05,
      "loss": 0.3733,
      "num_input_tokens_seen": 12954544,
      "step": 19815
    },
    {
      "epoch": 10.38784067085954,
      "grad_norm": 1.184253215789795,
      "learning_rate": 2.7668243904029894e-05,
      "loss": 0.5871,
      "num_input_tokens_seen": 12959568,
      "step": 19820
    },
    {
      "epoch": 10.390461215932914,
      "grad_norm": 0.3735866844654083,
      "learning_rate": 2.76568746522867e-05,
      "loss": 0.4561,
      "num_input_tokens_seen": 12962992,
      "step": 19825
    },
    {
      "epoch": 10.39308176100629,
      "grad_norm": 0.4497731328010559,
      "learning_rate": 2.7645504844755637e-05,
      "loss": 0.425,
      "num_input_tokens_seen": 12966608,
      "step": 19830
    },
    {
      "epoch": 10.395702306079665,
      "grad_norm": 0.2747673988342285,
      "learning_rate": 2.7634134483815143e-05,
      "loss": 0.4489,
      "num_input_tokens_seen": 12969744,
      "step": 19835
    },
    {
      "epoch": 10.39832285115304,
      "grad_norm": 0.7721583843231201,
      "learning_rate": 2.762276357184377e-05,
      "loss": 0.5323,
      "num_input_tokens_seen": 12972784,
      "step": 19840
    },
    {
      "epoch": 10.400943396226415,
      "grad_norm": 0.4107182025909424,
      "learning_rate": 2.7611392111220176e-05,
      "loss": 0.4331,
      "num_input_tokens_seen": 12976944,
      "step": 19845
    },
    {
      "epoch": 10.40356394129979,
      "grad_norm": 0.4023784101009369,
      "learning_rate": 2.7600020104323154e-05,
      "loss": 0.3365,
      "num_input_tokens_seen": 12979984,
      "step": 19850
    },
    {
      "epoch": 10.406184486373165,
      "grad_norm": 0.6686515212059021,
      "learning_rate": 2.7588647553531576e-05,
      "loss": 0.483,
      "num_input_tokens_seen": 12982704,
      "step": 19855
    },
    {
      "epoch": 10.40880503144654,
      "grad_norm": 0.4710981845855713,
      "learning_rate": 2.7577274461224463e-05,
      "loss": 0.4433,
      "num_input_tokens_seen": 12985680,
      "step": 19860
    },
    {
      "epoch": 10.411425576519916,
      "grad_norm": 0.445226788520813,
      "learning_rate": 2.756590082978095e-05,
      "loss": 0.4032,
      "num_input_tokens_seen": 12989200,
      "step": 19865
    },
    {
      "epoch": 10.414046121593291,
      "grad_norm": 0.2981571853160858,
      "learning_rate": 2.7554526661580243e-05,
      "loss": 0.4073,
      "num_input_tokens_seen": 12992560,
      "step": 19870
    },
    {
      "epoch": 10.416666666666666,
      "grad_norm": 0.593532145023346,
      "learning_rate": 2.7543151959001708e-05,
      "loss": 0.3792,
      "num_input_tokens_seen": 12995472,
      "step": 19875
    },
    {
      "epoch": 10.419287211740041,
      "grad_norm": 0.47264283895492554,
      "learning_rate": 2.7531776724424797e-05,
      "loss": 0.4805,
      "num_input_tokens_seen": 12999024,
      "step": 19880
    },
    {
      "epoch": 10.421907756813416,
      "grad_norm": 0.4213515520095825,
      "learning_rate": 2.752040096022907e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 13001744,
      "step": 19885
    },
    {
      "epoch": 10.424528301886792,
      "grad_norm": 0.6011782884597778,
      "learning_rate": 2.750902466879423e-05,
      "loss": 0.4907,
      "num_input_tokens_seen": 13004272,
      "step": 19890
    },
    {
      "epoch": 10.427148846960169,
      "grad_norm": 1.0493006706237793,
      "learning_rate": 2.7497647852500048e-05,
      "loss": 0.4875,
      "num_input_tokens_seen": 13007472,
      "step": 19895
    },
    {
      "epoch": 10.429769392033544,
      "grad_norm": 0.48014408349990845,
      "learning_rate": 2.7486270513726424e-05,
      "loss": 0.4238,
      "num_input_tokens_seen": 13011120,
      "step": 19900
    },
    {
      "epoch": 10.432389937106919,
      "grad_norm": 0.5680660009384155,
      "learning_rate": 2.7474892654853384e-05,
      "loss": 0.4367,
      "num_input_tokens_seen": 13013872,
      "step": 19905
    },
    {
      "epoch": 10.435010482180294,
      "grad_norm": 0.34588363766670227,
      "learning_rate": 2.7463514278261034e-05,
      "loss": 0.4723,
      "num_input_tokens_seen": 13016592,
      "step": 19910
    },
    {
      "epoch": 10.43763102725367,
      "grad_norm": 0.6713085174560547,
      "learning_rate": 2.7452135386329596e-05,
      "loss": 0.3618,
      "num_input_tokens_seen": 13019888,
      "step": 19915
    },
    {
      "epoch": 10.440251572327044,
      "grad_norm": 0.3465961515903473,
      "learning_rate": 2.7440755981439414e-05,
      "loss": 0.5776,
      "num_input_tokens_seen": 13023664,
      "step": 19920
    },
    {
      "epoch": 10.44287211740042,
      "grad_norm": 0.42236143350601196,
      "learning_rate": 2.7429376065970924e-05,
      "loss": 0.3772,
      "num_input_tokens_seen": 13027152,
      "step": 19925
    },
    {
      "epoch": 10.445492662473795,
      "grad_norm": 0.4495371878147125,
      "learning_rate": 2.7417995642304678e-05,
      "loss": 0.4614,
      "num_input_tokens_seen": 13030160,
      "step": 19930
    },
    {
      "epoch": 10.44811320754717,
      "grad_norm": 0.39797601103782654,
      "learning_rate": 2.7406614712821333e-05,
      "loss": 0.4119,
      "num_input_tokens_seen": 13034000,
      "step": 19935
    },
    {
      "epoch": 10.450733752620545,
      "grad_norm": 0.44886308908462524,
      "learning_rate": 2.739523327990164e-05,
      "loss": 0.3894,
      "num_input_tokens_seen": 13037392,
      "step": 19940
    },
    {
      "epoch": 10.45335429769392,
      "grad_norm": 0.646185040473938,
      "learning_rate": 2.7383851345926477e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 13040496,
      "step": 19945
    },
    {
      "epoch": 10.455974842767295,
      "grad_norm": 1.8753784894943237,
      "learning_rate": 2.7372468913276804e-05,
      "loss": 0.4277,
      "num_input_tokens_seen": 13043248,
      "step": 19950
    },
    {
      "epoch": 10.45859538784067,
      "grad_norm": 0.7609014511108398,
      "learning_rate": 2.7361085984333705e-05,
      "loss": 0.4205,
      "num_input_tokens_seen": 13046512,
      "step": 19955
    },
    {
      "epoch": 10.461215932914046,
      "grad_norm": 0.4573934078216553,
      "learning_rate": 2.7349702561478356e-05,
      "loss": 0.4329,
      "num_input_tokens_seen": 13049616,
      "step": 19960
    },
    {
      "epoch": 10.463836477987421,
      "grad_norm": 0.4993639588356018,
      "learning_rate": 2.733831864709203e-05,
      "loss": 0.3741,
      "num_input_tokens_seen": 13052048,
      "step": 19965
    },
    {
      "epoch": 10.466457023060796,
      "grad_norm": 0.8648959398269653,
      "learning_rate": 2.732693424355613e-05,
      "loss": 0.5378,
      "num_input_tokens_seen": 13054096,
      "step": 19970
    },
    {
      "epoch": 10.469077568134171,
      "grad_norm": 0.31135910749435425,
      "learning_rate": 2.731554935325213e-05,
      "loss": 0.3387,
      "num_input_tokens_seen": 13057488,
      "step": 19975
    },
    {
      "epoch": 10.471698113207546,
      "grad_norm": 0.35431888699531555,
      "learning_rate": 2.7304163978561613e-05,
      "loss": 0.3824,
      "num_input_tokens_seen": 13060336,
      "step": 19980
    },
    {
      "epoch": 10.474318658280922,
      "grad_norm": 0.411610871553421,
      "learning_rate": 2.729277812186629e-05,
      "loss": 0.4381,
      "num_input_tokens_seen": 13063728,
      "step": 19985
    },
    {
      "epoch": 10.476939203354299,
      "grad_norm": 0.42780447006225586,
      "learning_rate": 2.7281391785547945e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 13066256,
      "step": 19990
    },
    {
      "epoch": 10.479559748427674,
      "grad_norm": 0.6836908459663391,
      "learning_rate": 2.7270004971988456e-05,
      "loss": 0.5173,
      "num_input_tokens_seen": 13069424,
      "step": 19995
    },
    {
      "epoch": 10.482180293501049,
      "grad_norm": 0.49999406933784485,
      "learning_rate": 2.7258617683569825e-05,
      "loss": 0.3951,
      "num_input_tokens_seen": 13072560,
      "step": 20000
    },
    {
      "epoch": 10.484800838574424,
      "grad_norm": 0.494620680809021,
      "learning_rate": 2.7247229922674143e-05,
      "loss": 0.5561,
      "num_input_tokens_seen": 13076176,
      "step": 20005
    },
    {
      "epoch": 10.4874213836478,
      "grad_norm": 0.5239486694335938,
      "learning_rate": 2.7235841691683605e-05,
      "loss": 0.4258,
      "num_input_tokens_seen": 13079984,
      "step": 20010
    },
    {
      "epoch": 10.490041928721174,
      "grad_norm": 0.38236573338508606,
      "learning_rate": 2.7224452992980487e-05,
      "loss": 0.3938,
      "num_input_tokens_seen": 13082928,
      "step": 20015
    },
    {
      "epoch": 10.49266247379455,
      "grad_norm": 0.3200540542602539,
      "learning_rate": 2.7213063828947182e-05,
      "loss": 0.5373,
      "num_input_tokens_seen": 13085776,
      "step": 20020
    },
    {
      "epoch": 10.495283018867925,
      "grad_norm": 0.3175339996814728,
      "learning_rate": 2.7201674201966176e-05,
      "loss": 0.3872,
      "num_input_tokens_seen": 13088656,
      "step": 20025
    },
    {
      "epoch": 10.4979035639413,
      "grad_norm": 0.4938754141330719,
      "learning_rate": 2.7190284114420035e-05,
      "loss": 0.4652,
      "num_input_tokens_seen": 13091728,
      "step": 20030
    },
    {
      "epoch": 10.500524109014675,
      "grad_norm": 0.4308740496635437,
      "learning_rate": 2.717889356869146e-05,
      "loss": 0.5651,
      "num_input_tokens_seen": 13096048,
      "step": 20035
    },
    {
      "epoch": 10.50314465408805,
      "grad_norm": 0.43757811188697815,
      "learning_rate": 2.7167502567163194e-05,
      "loss": 0.5497,
      "num_input_tokens_seen": 13098608,
      "step": 20040
    },
    {
      "epoch": 10.505765199161425,
      "grad_norm": 0.4539405107498169,
      "learning_rate": 2.715611111221813e-05,
      "loss": 0.4409,
      "num_input_tokens_seen": 13101872,
      "step": 20045
    },
    {
      "epoch": 10.5083857442348,
      "grad_norm": 0.38313406705856323,
      "learning_rate": 2.7144719206239217e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 13109072,
      "step": 20050
    },
    {
      "epoch": 10.511006289308176,
      "grad_norm": 0.8255714774131775,
      "learning_rate": 2.7133326851609508e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 13112080,
      "step": 20055
    },
    {
      "epoch": 10.51362683438155,
      "grad_norm": 0.3666543960571289,
      "learning_rate": 2.7121934050712166e-05,
      "loss": 0.3629,
      "num_input_tokens_seen": 13114704,
      "step": 20060
    },
    {
      "epoch": 10.516247379454926,
      "grad_norm": 0.3501153290271759,
      "learning_rate": 2.711054080593043e-05,
      "loss": 0.5266,
      "num_input_tokens_seen": 13117360,
      "step": 20065
    },
    {
      "epoch": 10.518867924528301,
      "grad_norm": 0.37694084644317627,
      "learning_rate": 2.7099147119647627e-05,
      "loss": 0.4059,
      "num_input_tokens_seen": 13121104,
      "step": 20070
    },
    {
      "epoch": 10.521488469601676,
      "grad_norm": 0.4505425989627838,
      "learning_rate": 2.7087752994247206e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 13124464,
      "step": 20075
    },
    {
      "epoch": 10.524109014675052,
      "grad_norm": 0.45142194628715515,
      "learning_rate": 2.707635843211267e-05,
      "loss": 0.3605,
      "num_input_tokens_seen": 13127024,
      "step": 20080
    },
    {
      "epoch": 10.526729559748428,
      "grad_norm": 0.5165481567382812,
      "learning_rate": 2.7064963435627637e-05,
      "loss": 0.4235,
      "num_input_tokens_seen": 13129936,
      "step": 20085
    },
    {
      "epoch": 10.529350104821804,
      "grad_norm": 0.471754252910614,
      "learning_rate": 2.7053568007175823e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 13132656,
      "step": 20090
    },
    {
      "epoch": 10.531970649895179,
      "grad_norm": 0.45375320315361023,
      "learning_rate": 2.7042172149140994e-05,
      "loss": 0.4812,
      "num_input_tokens_seen": 13136528,
      "step": 20095
    },
    {
      "epoch": 10.534591194968554,
      "grad_norm": 0.47872406244277954,
      "learning_rate": 2.703077586390706e-05,
      "loss": 0.3939,
      "num_input_tokens_seen": 13139568,
      "step": 20100
    },
    {
      "epoch": 10.53721174004193,
      "grad_norm": 0.3910781443119049,
      "learning_rate": 2.7019379153857982e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 13142704,
      "step": 20105
    },
    {
      "epoch": 10.539832285115304,
      "grad_norm": 0.4389612078666687,
      "learning_rate": 2.700798202137782e-05,
      "loss": 0.5026,
      "num_input_tokens_seen": 13145936,
      "step": 20110
    },
    {
      "epoch": 10.54245283018868,
      "grad_norm": 0.4877811670303345,
      "learning_rate": 2.6996584468850732e-05,
      "loss": 0.5423,
      "num_input_tokens_seen": 13149360,
      "step": 20115
    },
    {
      "epoch": 10.545073375262055,
      "grad_norm": 0.7995206713676453,
      "learning_rate": 2.6985186498660946e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 13151920,
      "step": 20120
    },
    {
      "epoch": 10.54769392033543,
      "grad_norm": 0.3710925579071045,
      "learning_rate": 2.69737881131928e-05,
      "loss": 0.4941,
      "num_input_tokens_seen": 13155024,
      "step": 20125
    },
    {
      "epoch": 10.550314465408805,
      "grad_norm": 0.4449176788330078,
      "learning_rate": 2.6962389314830692e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 13158320,
      "step": 20130
    },
    {
      "epoch": 10.55293501048218,
      "grad_norm": 0.5318021178245544,
      "learning_rate": 2.6950990105959135e-05,
      "loss": 0.4262,
      "num_input_tokens_seen": 13161040,
      "step": 20135
    },
    {
      "epoch": 10.555555555555555,
      "grad_norm": 0.9187355041503906,
      "learning_rate": 2.6939590488962703e-05,
      "loss": 0.3733,
      "num_input_tokens_seen": 13163984,
      "step": 20140
    },
    {
      "epoch": 10.55817610062893,
      "grad_norm": 0.49084413051605225,
      "learning_rate": 2.6928190466226067e-05,
      "loss": 0.5468,
      "num_input_tokens_seen": 13167888,
      "step": 20145
    },
    {
      "epoch": 10.560796645702306,
      "grad_norm": 0.2768601179122925,
      "learning_rate": 2.691679004013398e-05,
      "loss": 0.3972,
      "num_input_tokens_seen": 13171312,
      "step": 20150
    },
    {
      "epoch": 10.56341719077568,
      "grad_norm": 0.5276437401771545,
      "learning_rate": 2.6905389213071285e-05,
      "loss": 0.4578,
      "num_input_tokens_seen": 13174192,
      "step": 20155
    },
    {
      "epoch": 10.566037735849056,
      "grad_norm": 0.33495256304740906,
      "learning_rate": 2.6893987987422907e-05,
      "loss": 0.5278,
      "num_input_tokens_seen": 13178736,
      "step": 20160
    },
    {
      "epoch": 10.568658280922431,
      "grad_norm": 0.3926340341567993,
      "learning_rate": 2.6882586365573854e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 13181392,
      "step": 20165
    },
    {
      "epoch": 10.571278825995806,
      "grad_norm": 0.3740251958370209,
      "learning_rate": 2.6871184349909206e-05,
      "loss": 0.3835,
      "num_input_tokens_seen": 13184496,
      "step": 20170
    },
    {
      "epoch": 10.573899371069182,
      "grad_norm": 0.552017331123352,
      "learning_rate": 2.685978194281414e-05,
      "loss": 0.5183,
      "num_input_tokens_seen": 13187760,
      "step": 20175
    },
    {
      "epoch": 10.576519916142558,
      "grad_norm": 0.62831050157547,
      "learning_rate": 2.6848379146673914e-05,
      "loss": 0.3673,
      "num_input_tokens_seen": 13190640,
      "step": 20180
    },
    {
      "epoch": 10.579140461215934,
      "grad_norm": 0.615987241268158,
      "learning_rate": 2.6836975963873843e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 13193840,
      "step": 20185
    },
    {
      "epoch": 10.581761006289309,
      "grad_norm": 0.5050910711288452,
      "learning_rate": 2.6825572396799377e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 13197648,
      "step": 20190
    },
    {
      "epoch": 10.584381551362684,
      "grad_norm": 0.46820318698883057,
      "learning_rate": 2.6814168447835986e-05,
      "loss": 0.2688,
      "num_input_tokens_seen": 13201808,
      "step": 20195
    },
    {
      "epoch": 10.58700209643606,
      "grad_norm": 0.5348331928253174,
      "learning_rate": 2.6802764119369245e-05,
      "loss": 0.5642,
      "num_input_tokens_seen": 13204976,
      "step": 20200
    },
    {
      "epoch": 10.589622641509434,
      "grad_norm": 0.6126003265380859,
      "learning_rate": 2.6791359413784828e-05,
      "loss": 0.5412,
      "num_input_tokens_seen": 13207856,
      "step": 20205
    },
    {
      "epoch": 10.59224318658281,
      "grad_norm": 0.8741165995597839,
      "learning_rate": 2.677995433346845e-05,
      "loss": 0.3497,
      "num_input_tokens_seen": 13210768,
      "step": 20210
    },
    {
      "epoch": 10.594863731656185,
      "grad_norm": 0.6109898686408997,
      "learning_rate": 2.6768548880805934e-05,
      "loss": 0.5034,
      "num_input_tokens_seen": 13213776,
      "step": 20215
    },
    {
      "epoch": 10.59748427672956,
      "grad_norm": 0.34141138195991516,
      "learning_rate": 2.6757143058183177e-05,
      "loss": 0.5823,
      "num_input_tokens_seen": 13217104,
      "step": 20220
    },
    {
      "epoch": 10.600104821802935,
      "grad_norm": 0.25424471497535706,
      "learning_rate": 2.6745736867986132e-05,
      "loss": 0.3918,
      "num_input_tokens_seen": 13220048,
      "step": 20225
    },
    {
      "epoch": 10.60272536687631,
      "grad_norm": 0.5738143920898438,
      "learning_rate": 2.673433031260085e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 13223376,
      "step": 20230
    },
    {
      "epoch": 10.605345911949685,
      "grad_norm": 0.44250884652137756,
      "learning_rate": 2.6722923394413456e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 13226512,
      "step": 20235
    },
    {
      "epoch": 10.60796645702306,
      "grad_norm": 0.29955875873565674,
      "learning_rate": 2.6711516115810136e-05,
      "loss": 0.3734,
      "num_input_tokens_seen": 13229744,
      "step": 20240
    },
    {
      "epoch": 10.610587002096436,
      "grad_norm": 0.34402695298194885,
      "learning_rate": 2.6700108479177188e-05,
      "loss": 0.4327,
      "num_input_tokens_seen": 13233552,
      "step": 20245
    },
    {
      "epoch": 10.61320754716981,
      "grad_norm": 0.5416373014450073,
      "learning_rate": 2.6688700486900926e-05,
      "loss": 0.3854,
      "num_input_tokens_seen": 13237008,
      "step": 20250
    },
    {
      "epoch": 10.615828092243186,
      "grad_norm": 0.46077123284339905,
      "learning_rate": 2.6677292141367787e-05,
      "loss": 0.3115,
      "num_input_tokens_seen": 13240080,
      "step": 20255
    },
    {
      "epoch": 10.618448637316561,
      "grad_norm": 0.5633283257484436,
      "learning_rate": 2.6665883444964278e-05,
      "loss": 0.4618,
      "num_input_tokens_seen": 13243408,
      "step": 20260
    },
    {
      "epoch": 10.621069182389936,
      "grad_norm": 0.44445204734802246,
      "learning_rate": 2.665447440007695e-05,
      "loss": 0.5208,
      "num_input_tokens_seen": 13247152,
      "step": 20265
    },
    {
      "epoch": 10.623689727463312,
      "grad_norm": 0.32526281476020813,
      "learning_rate": 2.664306500909246e-05,
      "loss": 0.313,
      "num_input_tokens_seen": 13251376,
      "step": 20270
    },
    {
      "epoch": 10.626310272536688,
      "grad_norm": 0.3879692554473877,
      "learning_rate": 2.663165527439751e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 13254992,
      "step": 20275
    },
    {
      "epoch": 10.628930817610064,
      "grad_norm": 0.40960177779197693,
      "learning_rate": 2.662024519837889e-05,
      "loss": 0.4296,
      "num_input_tokens_seen": 13257712,
      "step": 20280
    },
    {
      "epoch": 10.631551362683439,
      "grad_norm": 0.5270946621894836,
      "learning_rate": 2.6608834783423465e-05,
      "loss": 0.3743,
      "num_input_tokens_seen": 13262192,
      "step": 20285
    },
    {
      "epoch": 10.634171907756814,
      "grad_norm": 0.3036913275718689,
      "learning_rate": 2.6597424031918156e-05,
      "loss": 0.489,
      "num_input_tokens_seen": 13266832,
      "step": 20290
    },
    {
      "epoch": 10.63679245283019,
      "grad_norm": 0.7297710180282593,
      "learning_rate": 2.6586012946249954e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 13269424,
      "step": 20295
    },
    {
      "epoch": 10.639412997903564,
      "grad_norm": 0.6187723875045776,
      "learning_rate": 2.6574601528805953e-05,
      "loss": 0.5851,
      "num_input_tokens_seen": 13272144,
      "step": 20300
    },
    {
      "epoch": 10.64203354297694,
      "grad_norm": 0.45535847544670105,
      "learning_rate": 2.6563189781973268e-05,
      "loss": 0.4094,
      "num_input_tokens_seen": 13275376,
      "step": 20305
    },
    {
      "epoch": 10.644654088050315,
      "grad_norm": 0.35549989342689514,
      "learning_rate": 2.655177770813912e-05,
      "loss": 0.3979,
      "num_input_tokens_seen": 13278320,
      "step": 20310
    },
    {
      "epoch": 10.64727463312369,
      "grad_norm": 0.858362078666687,
      "learning_rate": 2.654036530969077e-05,
      "loss": 0.4852,
      "num_input_tokens_seen": 13280816,
      "step": 20315
    },
    {
      "epoch": 10.649895178197065,
      "grad_norm": 0.45773786306381226,
      "learning_rate": 2.6528952589015576e-05,
      "loss": 0.6413,
      "num_input_tokens_seen": 13284784,
      "step": 20320
    },
    {
      "epoch": 10.65251572327044,
      "grad_norm": 0.4533621668815613,
      "learning_rate": 2.6517539548500947e-05,
      "loss": 0.5891,
      "num_input_tokens_seen": 13287376,
      "step": 20325
    },
    {
      "epoch": 10.655136268343815,
      "grad_norm": 0.34576165676116943,
      "learning_rate": 2.6506126190534354e-05,
      "loss": 0.2976,
      "num_input_tokens_seen": 13290928,
      "step": 20330
    },
    {
      "epoch": 10.65775681341719,
      "grad_norm": 0.8586809039115906,
      "learning_rate": 2.649471251750334e-05,
      "loss": 0.5239,
      "num_input_tokens_seen": 13294192,
      "step": 20335
    },
    {
      "epoch": 10.660377358490566,
      "grad_norm": 0.6041765809059143,
      "learning_rate": 2.6483298531795527e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 13297456,
      "step": 20340
    },
    {
      "epoch": 10.66299790356394,
      "grad_norm": 0.521037220954895,
      "learning_rate": 2.6471884235798583e-05,
      "loss": 0.6179,
      "num_input_tokens_seen": 13301008,
      "step": 20345
    },
    {
      "epoch": 10.665618448637316,
      "grad_norm": 0.7033892273902893,
      "learning_rate": 2.6460469631900247e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 13304368,
      "step": 20350
    },
    {
      "epoch": 10.668238993710691,
      "grad_norm": 0.4525379538536072,
      "learning_rate": 2.644905472248832e-05,
      "loss": 0.5559,
      "num_input_tokens_seen": 13306864,
      "step": 20355
    },
    {
      "epoch": 10.670859538784066,
      "grad_norm": 0.5311416387557983,
      "learning_rate": 2.6437639509950686e-05,
      "loss": 0.4342,
      "num_input_tokens_seen": 13310352,
      "step": 20360
    },
    {
      "epoch": 10.673480083857442,
      "grad_norm": 0.6620708107948303,
      "learning_rate": 2.6426223996675264e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 13313488,
      "step": 20365
    },
    {
      "epoch": 10.676100628930818,
      "grad_norm": 0.7437820434570312,
      "learning_rate": 2.641480818505005e-05,
      "loss": 0.4253,
      "num_input_tokens_seen": 13316112,
      "step": 20370
    },
    {
      "epoch": 10.678721174004194,
      "grad_norm": 0.3230387568473816,
      "learning_rate": 2.6403392077463108e-05,
      "loss": 0.4361,
      "num_input_tokens_seen": 13320528,
      "step": 20375
    },
    {
      "epoch": 10.681341719077569,
      "grad_norm": 0.35214149951934814,
      "learning_rate": 2.639197567630256e-05,
      "loss": 0.466,
      "num_input_tokens_seen": 13323888,
      "step": 20380
    },
    {
      "epoch": 10.683962264150944,
      "grad_norm": 0.5509995222091675,
      "learning_rate": 2.638055898395657e-05,
      "loss": 0.5398,
      "num_input_tokens_seen": 13327568,
      "step": 20385
    },
    {
      "epoch": 10.68658280922432,
      "grad_norm": 0.7658198475837708,
      "learning_rate": 2.6369142002813403e-05,
      "loss": 0.547,
      "num_input_tokens_seen": 13330096,
      "step": 20390
    },
    {
      "epoch": 10.689203354297694,
      "grad_norm": 0.6601452231407166,
      "learning_rate": 2.6357724735261335e-05,
      "loss": 0.4144,
      "num_input_tokens_seen": 13332560,
      "step": 20395
    },
    {
      "epoch": 10.69182389937107,
      "grad_norm": 0.437395840883255,
      "learning_rate": 2.6346307183688752e-05,
      "loss": 0.5674,
      "num_input_tokens_seen": 13336080,
      "step": 20400
    },
    {
      "epoch": 10.694444444444445,
      "grad_norm": 0.8142441511154175,
      "learning_rate": 2.6334889350484064e-05,
      "loss": 0.4051,
      "num_input_tokens_seen": 13339696,
      "step": 20405
    },
    {
      "epoch": 10.69706498951782,
      "grad_norm": 0.3137974441051483,
      "learning_rate": 2.632347123803575e-05,
      "loss": 0.413,
      "num_input_tokens_seen": 13343760,
      "step": 20410
    },
    {
      "epoch": 10.699685534591195,
      "grad_norm": 0.5297717452049255,
      "learning_rate": 2.6312052848732353e-05,
      "loss": 0.4973,
      "num_input_tokens_seen": 13346928,
      "step": 20415
    },
    {
      "epoch": 10.70230607966457,
      "grad_norm": 0.43564918637275696,
      "learning_rate": 2.6300634184962475e-05,
      "loss": 0.5355,
      "num_input_tokens_seen": 13349968,
      "step": 20420
    },
    {
      "epoch": 10.704926624737945,
      "grad_norm": 0.8227455615997314,
      "learning_rate": 2.628921524911475e-05,
      "loss": 0.4791,
      "num_input_tokens_seen": 13353232,
      "step": 20425
    },
    {
      "epoch": 10.70754716981132,
      "grad_norm": 0.5820368528366089,
      "learning_rate": 2.6277796043577924e-05,
      "loss": 0.4322,
      "num_input_tokens_seen": 13356496,
      "step": 20430
    },
    {
      "epoch": 10.710167714884696,
      "grad_norm": 0.455087810754776,
      "learning_rate": 2.6266376570740725e-05,
      "loss": 0.4688,
      "num_input_tokens_seen": 13360656,
      "step": 20435
    },
    {
      "epoch": 10.71278825995807,
      "grad_norm": 0.41292667388916016,
      "learning_rate": 2.6254956832992005e-05,
      "loss": 0.3517,
      "num_input_tokens_seen": 13363984,
      "step": 20440
    },
    {
      "epoch": 10.715408805031446,
      "grad_norm": 0.737895667552948,
      "learning_rate": 2.6243536832720627e-05,
      "loss": 0.4689,
      "num_input_tokens_seen": 13367664,
      "step": 20445
    },
    {
      "epoch": 10.718029350104821,
      "grad_norm": 1.0577151775360107,
      "learning_rate": 2.623211657231553e-05,
      "loss": 0.3503,
      "num_input_tokens_seen": 13369744,
      "step": 20450
    },
    {
      "epoch": 10.720649895178196,
      "grad_norm": 0.4409232437610626,
      "learning_rate": 2.6220696054165704e-05,
      "loss": 0.4373,
      "num_input_tokens_seen": 13373584,
      "step": 20455
    },
    {
      "epoch": 10.723270440251572,
      "grad_norm": 0.3399665653705597,
      "learning_rate": 2.6209275280660194e-05,
      "loss": 0.4968,
      "num_input_tokens_seen": 13377232,
      "step": 20460
    },
    {
      "epoch": 10.725890985324948,
      "grad_norm": 0.4704323410987854,
      "learning_rate": 2.6197854254188085e-05,
      "loss": 0.5765,
      "num_input_tokens_seen": 13380944,
      "step": 20465
    },
    {
      "epoch": 10.728511530398324,
      "grad_norm": 0.6475104689598083,
      "learning_rate": 2.618643297713854e-05,
      "loss": 0.4525,
      "num_input_tokens_seen": 13385072,
      "step": 20470
    },
    {
      "epoch": 10.731132075471699,
      "grad_norm": 0.46064072847366333,
      "learning_rate": 2.617501145190075e-05,
      "loss": 0.4424,
      "num_input_tokens_seen": 13388016,
      "step": 20475
    },
    {
      "epoch": 10.733752620545074,
      "grad_norm": 0.3705517649650574,
      "learning_rate": 2.616358968086397e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 13391760,
      "step": 20480
    },
    {
      "epoch": 10.73637316561845,
      "grad_norm": 0.4398510158061981,
      "learning_rate": 2.6152167666417493e-05,
      "loss": 0.5243,
      "num_input_tokens_seen": 13396144,
      "step": 20485
    },
    {
      "epoch": 10.738993710691824,
      "grad_norm": 0.654845118522644,
      "learning_rate": 2.6140745410950694e-05,
      "loss": 0.3654,
      "num_input_tokens_seen": 13399824,
      "step": 20490
    },
    {
      "epoch": 10.7416142557652,
      "grad_norm": 0.47253209352493286,
      "learning_rate": 2.6129322916852973e-05,
      "loss": 0.4162,
      "num_input_tokens_seen": 13403248,
      "step": 20495
    },
    {
      "epoch": 10.744234800838575,
      "grad_norm": 0.5844261050224304,
      "learning_rate": 2.611790018651377e-05,
      "loss": 0.3959,
      "num_input_tokens_seen": 13405904,
      "step": 20500
    },
    {
      "epoch": 10.74685534591195,
      "grad_norm": 0.5606511831283569,
      "learning_rate": 2.6106477222322613e-05,
      "loss": 0.3598,
      "num_input_tokens_seen": 13409360,
      "step": 20505
    },
    {
      "epoch": 10.749475890985325,
      "grad_norm": 0.4161536693572998,
      "learning_rate": 2.6095054026669037e-05,
      "loss": 0.4177,
      "num_input_tokens_seen": 13412304,
      "step": 20510
    },
    {
      "epoch": 10.7520964360587,
      "grad_norm": 0.5425213575363159,
      "learning_rate": 2.608363060194265e-05,
      "loss": 0.4679,
      "num_input_tokens_seen": 13415568,
      "step": 20515
    },
    {
      "epoch": 10.754716981132075,
      "grad_norm": 0.5109556317329407,
      "learning_rate": 2.6072206950533117e-05,
      "loss": 0.4774,
      "num_input_tokens_seen": 13418160,
      "step": 20520
    },
    {
      "epoch": 10.75733752620545,
      "grad_norm": 0.5599351525306702,
      "learning_rate": 2.606078307483011e-05,
      "loss": 0.3221,
      "num_input_tokens_seen": 13421168,
      "step": 20525
    },
    {
      "epoch": 10.759958071278826,
      "grad_norm": 0.38795843720436096,
      "learning_rate": 2.6049358977223392e-05,
      "loss": 0.3985,
      "num_input_tokens_seen": 13423824,
      "step": 20530
    },
    {
      "epoch": 10.7625786163522,
      "grad_norm": 0.480542927980423,
      "learning_rate": 2.6037934660102747e-05,
      "loss": 0.3871,
      "num_input_tokens_seen": 13426928,
      "step": 20535
    },
    {
      "epoch": 10.765199161425576,
      "grad_norm": 0.4037693738937378,
      "learning_rate": 2.6026510125858006e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 13429712,
      "step": 20540
    },
    {
      "epoch": 10.767819706498951,
      "grad_norm": 0.310479074716568,
      "learning_rate": 2.6015085376879063e-05,
      "loss": 0.3641,
      "num_input_tokens_seen": 13433648,
      "step": 20545
    },
    {
      "epoch": 10.770440251572326,
      "grad_norm": 0.5244796276092529,
      "learning_rate": 2.600366041555584e-05,
      "loss": 0.403,
      "num_input_tokens_seen": 13436048,
      "step": 20550
    },
    {
      "epoch": 10.773060796645701,
      "grad_norm": 0.5246856808662415,
      "learning_rate": 2.5992235244278302e-05,
      "loss": 0.4796,
      "num_input_tokens_seen": 13438672,
      "step": 20555
    },
    {
      "epoch": 10.775681341719078,
      "grad_norm": 0.4401644170284271,
      "learning_rate": 2.5980809865436485e-05,
      "loss": 0.366,
      "num_input_tokens_seen": 13440976,
      "step": 20560
    },
    {
      "epoch": 10.778301886792454,
      "grad_norm": 0.3905618190765381,
      "learning_rate": 2.5969384281420424e-05,
      "loss": 0.3421,
      "num_input_tokens_seen": 13444528,
      "step": 20565
    },
    {
      "epoch": 10.780922431865829,
      "grad_norm": 1.0419507026672363,
      "learning_rate": 2.5957958494620227e-05,
      "loss": 0.4972,
      "num_input_tokens_seen": 13447664,
      "step": 20570
    },
    {
      "epoch": 10.783542976939204,
      "grad_norm": 0.3432142734527588,
      "learning_rate": 2.594653250742605e-05,
      "loss": 0.5038,
      "num_input_tokens_seen": 13451312,
      "step": 20575
    },
    {
      "epoch": 10.786163522012579,
      "grad_norm": 0.4204134941101074,
      "learning_rate": 2.5935106322228065e-05,
      "loss": 0.4551,
      "num_input_tokens_seen": 13453776,
      "step": 20580
    },
    {
      "epoch": 10.788784067085954,
      "grad_norm": 0.7368908524513245,
      "learning_rate": 2.5923679941416514e-05,
      "loss": 0.4229,
      "num_input_tokens_seen": 13456656,
      "step": 20585
    },
    {
      "epoch": 10.79140461215933,
      "grad_norm": 0.44665592908859253,
      "learning_rate": 2.591225336738165e-05,
      "loss": 0.4352,
      "num_input_tokens_seen": 13459440,
      "step": 20590
    },
    {
      "epoch": 10.794025157232705,
      "grad_norm": 0.4444838762283325,
      "learning_rate": 2.5900826602513782e-05,
      "loss": 0.5375,
      "num_input_tokens_seen": 13462608,
      "step": 20595
    },
    {
      "epoch": 10.79664570230608,
      "grad_norm": 0.4943983852863312,
      "learning_rate": 2.5889399649203277e-05,
      "loss": 0.4283,
      "num_input_tokens_seen": 13465424,
      "step": 20600
    },
    {
      "epoch": 10.799266247379455,
      "grad_norm": 0.6480081677436829,
      "learning_rate": 2.5877972509840514e-05,
      "loss": 0.4147,
      "num_input_tokens_seen": 13468688,
      "step": 20605
    },
    {
      "epoch": 10.80188679245283,
      "grad_norm": 0.56378173828125,
      "learning_rate": 2.5866545186815912e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 13471312,
      "step": 20610
    },
    {
      "epoch": 10.804507337526205,
      "grad_norm": 0.8795725703239441,
      "learning_rate": 2.5855117682519954e-05,
      "loss": 0.2775,
      "num_input_tokens_seen": 13474352,
      "step": 20615
    },
    {
      "epoch": 10.80712788259958,
      "grad_norm": 0.380740761756897,
      "learning_rate": 2.584368999934312e-05,
      "loss": 0.5113,
      "num_input_tokens_seen": 13478704,
      "step": 20620
    },
    {
      "epoch": 10.809748427672956,
      "grad_norm": 0.6797909736633301,
      "learning_rate": 2.5832262139675966e-05,
      "loss": 0.4354,
      "num_input_tokens_seen": 13481904,
      "step": 20625
    },
    {
      "epoch": 10.81236897274633,
      "grad_norm": 0.5340741872787476,
      "learning_rate": 2.582083410590907e-05,
      "loss": 0.5961,
      "num_input_tokens_seen": 13484624,
      "step": 20630
    },
    {
      "epoch": 10.814989517819706,
      "grad_norm": 0.34499430656433105,
      "learning_rate": 2.5809405900433036e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 13488368,
      "step": 20635
    },
    {
      "epoch": 10.817610062893081,
      "grad_norm": 0.4647369682788849,
      "learning_rate": 2.579797752563853e-05,
      "loss": 0.5805,
      "num_input_tokens_seen": 13490672,
      "step": 20640
    },
    {
      "epoch": 10.820230607966456,
      "grad_norm": 0.560956597328186,
      "learning_rate": 2.5786548983916232e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 13495984,
      "step": 20645
    },
    {
      "epoch": 10.822851153039831,
      "grad_norm": 0.5552095174789429,
      "learning_rate": 2.577512027765685e-05,
      "loss": 0.4286,
      "num_input_tokens_seen": 13498416,
      "step": 20650
    },
    {
      "epoch": 10.825471698113208,
      "grad_norm": 0.31380459666252136,
      "learning_rate": 2.5763691409251155e-05,
      "loss": 0.4904,
      "num_input_tokens_seen": 13502512,
      "step": 20655
    },
    {
      "epoch": 10.828092243186584,
      "grad_norm": 0.3076685965061188,
      "learning_rate": 2.5752262381089932e-05,
      "loss": 0.491,
      "num_input_tokens_seen": 13506416,
      "step": 20660
    },
    {
      "epoch": 10.830712788259959,
      "grad_norm": 0.5102716088294983,
      "learning_rate": 2.5740833195563996e-05,
      "loss": 0.5748,
      "num_input_tokens_seen": 13509392,
      "step": 20665
    },
    {
      "epoch": 10.833333333333334,
      "grad_norm": 1.1555660963058472,
      "learning_rate": 2.572940385506421e-05,
      "loss": 0.522,
      "num_input_tokens_seen": 13514032,
      "step": 20670
    },
    {
      "epoch": 10.835953878406709,
      "grad_norm": 0.5032921433448792,
      "learning_rate": 2.5717974361981457e-05,
      "loss": 0.4974,
      "num_input_tokens_seen": 13517968,
      "step": 20675
    },
    {
      "epoch": 10.838574423480084,
      "grad_norm": 0.24897295236587524,
      "learning_rate": 2.5706544718706665e-05,
      "loss": 0.5207,
      "num_input_tokens_seen": 13521616,
      "step": 20680
    },
    {
      "epoch": 10.84119496855346,
      "grad_norm": 0.41666504740715027,
      "learning_rate": 2.569511492763077e-05,
      "loss": 0.3103,
      "num_input_tokens_seen": 13524592,
      "step": 20685
    },
    {
      "epoch": 10.843815513626835,
      "grad_norm": 0.555317759513855,
      "learning_rate": 2.5683684991144764e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 13528048,
      "step": 20690
    },
    {
      "epoch": 10.84643605870021,
      "grad_norm": 0.4269776940345764,
      "learning_rate": 2.5672254911639666e-05,
      "loss": 0.4514,
      "num_input_tokens_seen": 13531120,
      "step": 20695
    },
    {
      "epoch": 10.849056603773585,
      "grad_norm": 0.5283628106117249,
      "learning_rate": 2.5660824691506496e-05,
      "loss": 0.4841,
      "num_input_tokens_seen": 13534288,
      "step": 20700
    },
    {
      "epoch": 10.85167714884696,
      "grad_norm": 0.29478245973587036,
      "learning_rate": 2.5649394333136356e-05,
      "loss": 0.5965,
      "num_input_tokens_seen": 13538032,
      "step": 20705
    },
    {
      "epoch": 10.854297693920335,
      "grad_norm": 0.5974480509757996,
      "learning_rate": 2.5637963838920316e-05,
      "loss": 0.4672,
      "num_input_tokens_seen": 13541008,
      "step": 20710
    },
    {
      "epoch": 10.85691823899371,
      "grad_norm": 0.6379504799842834,
      "learning_rate": 2.5626533211249525e-05,
      "loss": 0.4911,
      "num_input_tokens_seen": 13544304,
      "step": 20715
    },
    {
      "epoch": 10.859538784067086,
      "grad_norm": 0.44687631726264954,
      "learning_rate": 2.5615102452515138e-05,
      "loss": 0.49,
      "num_input_tokens_seen": 13548048,
      "step": 20720
    },
    {
      "epoch": 10.86215932914046,
      "grad_norm": 0.42049771547317505,
      "learning_rate": 2.5603671565108328e-05,
      "loss": 0.3231,
      "num_input_tokens_seen": 13554288,
      "step": 20725
    },
    {
      "epoch": 10.864779874213836,
      "grad_norm": 0.5990914106369019,
      "learning_rate": 2.5592240551420317e-05,
      "loss": 0.5201,
      "num_input_tokens_seen": 13558320,
      "step": 20730
    },
    {
      "epoch": 10.867400419287211,
      "grad_norm": 0.6317090392112732,
      "learning_rate": 2.5580809413842343e-05,
      "loss": 0.3617,
      "num_input_tokens_seen": 13561328,
      "step": 20735
    },
    {
      "epoch": 10.870020964360586,
      "grad_norm": 0.5084317326545715,
      "learning_rate": 2.5569378154765655e-05,
      "loss": 0.5727,
      "num_input_tokens_seen": 13566224,
      "step": 20740
    },
    {
      "epoch": 10.872641509433961,
      "grad_norm": 0.8659859895706177,
      "learning_rate": 2.5557946776581572e-05,
      "loss": 0.5828,
      "num_input_tokens_seen": 13568752,
      "step": 20745
    },
    {
      "epoch": 10.875262054507338,
      "grad_norm": 0.4037097990512848,
      "learning_rate": 2.554651528168137e-05,
      "loss": 0.3909,
      "num_input_tokens_seen": 13571664,
      "step": 20750
    },
    {
      "epoch": 10.877882599580714,
      "grad_norm": 0.35952916741371155,
      "learning_rate": 2.5535083672456407e-05,
      "loss": 0.4509,
      "num_input_tokens_seen": 13574704,
      "step": 20755
    },
    {
      "epoch": 10.880503144654089,
      "grad_norm": 0.35223615169525146,
      "learning_rate": 2.5523651951298055e-05,
      "loss": 0.4861,
      "num_input_tokens_seen": 13577264,
      "step": 20760
    },
    {
      "epoch": 10.883123689727464,
      "grad_norm": 0.6474040150642395,
      "learning_rate": 2.5512220120597686e-05,
      "loss": 0.3659,
      "num_input_tokens_seen": 13580240,
      "step": 20765
    },
    {
      "epoch": 10.885744234800839,
      "grad_norm": 0.29104167222976685,
      "learning_rate": 2.550078818274671e-05,
      "loss": 0.3895,
      "num_input_tokens_seen": 13583472,
      "step": 20770
    },
    {
      "epoch": 10.888364779874214,
      "grad_norm": 0.40818142890930176,
      "learning_rate": 2.548935614013656e-05,
      "loss": 0.4385,
      "num_input_tokens_seen": 13586640,
      "step": 20775
    },
    {
      "epoch": 10.89098532494759,
      "grad_norm": 0.49033451080322266,
      "learning_rate": 2.5477923995158683e-05,
      "loss": 0.445,
      "num_input_tokens_seen": 13589680,
      "step": 20780
    },
    {
      "epoch": 10.893605870020965,
      "grad_norm": 0.8842618465423584,
      "learning_rate": 2.5466491750204562e-05,
      "loss": 0.6443,
      "num_input_tokens_seen": 13593072,
      "step": 20785
    },
    {
      "epoch": 10.89622641509434,
      "grad_norm": 0.4014686048030853,
      "learning_rate": 2.5455059407665693e-05,
      "loss": 0.4058,
      "num_input_tokens_seen": 13596624,
      "step": 20790
    },
    {
      "epoch": 10.898846960167715,
      "grad_norm": 0.3058764636516571,
      "learning_rate": 2.544362696993358e-05,
      "loss": 0.3804,
      "num_input_tokens_seen": 13599088,
      "step": 20795
    },
    {
      "epoch": 10.90146750524109,
      "grad_norm": 0.5263655185699463,
      "learning_rate": 2.5432194439399775e-05,
      "loss": 0.4082,
      "num_input_tokens_seen": 13601552,
      "step": 20800
    },
    {
      "epoch": 10.904088050314465,
      "grad_norm": 0.5088794231414795,
      "learning_rate": 2.542076181845582e-05,
      "loss": 0.3765,
      "num_input_tokens_seen": 13604144,
      "step": 20805
    },
    {
      "epoch": 10.90670859538784,
      "grad_norm": 0.8169078826904297,
      "learning_rate": 2.540932910949329e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 13606928,
      "step": 20810
    },
    {
      "epoch": 10.909329140461216,
      "grad_norm": 0.4773993492126465,
      "learning_rate": 2.5397896314903785e-05,
      "loss": 0.4245,
      "num_input_tokens_seen": 13610320,
      "step": 20815
    },
    {
      "epoch": 10.91194968553459,
      "grad_norm": 0.18411938846111298,
      "learning_rate": 2.5386463437078905e-05,
      "loss": 0.6829,
      "num_input_tokens_seen": 13613680,
      "step": 20820
    },
    {
      "epoch": 10.914570230607966,
      "grad_norm": 0.6715101003646851,
      "learning_rate": 2.5375030478410282e-05,
      "loss": 0.3624,
      "num_input_tokens_seen": 13617264,
      "step": 20825
    },
    {
      "epoch": 10.917190775681341,
      "grad_norm": 0.540006697177887,
      "learning_rate": 2.536359744128957e-05,
      "loss": 0.426,
      "num_input_tokens_seen": 13620208,
      "step": 20830
    },
    {
      "epoch": 10.919811320754716,
      "grad_norm": 0.5255924463272095,
      "learning_rate": 2.5352164328108413e-05,
      "loss": 0.408,
      "num_input_tokens_seen": 13623248,
      "step": 20835
    },
    {
      "epoch": 10.922431865828091,
      "grad_norm": 0.5389805436134338,
      "learning_rate": 2.53407311412585e-05,
      "loss": 0.4278,
      "num_input_tokens_seen": 13626672,
      "step": 20840
    },
    {
      "epoch": 10.925052410901468,
      "grad_norm": 0.3257617652416229,
      "learning_rate": 2.5329297883131526e-05,
      "loss": 0.4768,
      "num_input_tokens_seen": 13629776,
      "step": 20845
    },
    {
      "epoch": 10.927672955974844,
      "grad_norm": 0.4609209895133972,
      "learning_rate": 2.5317864556119182e-05,
      "loss": 0.5072,
      "num_input_tokens_seen": 13632336,
      "step": 20850
    },
    {
      "epoch": 10.930293501048219,
      "grad_norm": 0.7618720531463623,
      "learning_rate": 2.53064311626132e-05,
      "loss": 0.4954,
      "num_input_tokens_seen": 13635600,
      "step": 20855
    },
    {
      "epoch": 10.932914046121594,
      "grad_norm": 0.5094859004020691,
      "learning_rate": 2.5294997705005318e-05,
      "loss": 0.4148,
      "num_input_tokens_seen": 13639248,
      "step": 20860
    },
    {
      "epoch": 10.935534591194969,
      "grad_norm": 0.7231926918029785,
      "learning_rate": 2.5283564185687282e-05,
      "loss": 0.4007,
      "num_input_tokens_seen": 13641872,
      "step": 20865
    },
    {
      "epoch": 10.938155136268344,
      "grad_norm": 0.3690437376499176,
      "learning_rate": 2.5272130607050848e-05,
      "loss": 0.406,
      "num_input_tokens_seen": 13645424,
      "step": 20870
    },
    {
      "epoch": 10.94077568134172,
      "grad_norm": 0.29850178956985474,
      "learning_rate": 2.5260696971487802e-05,
      "loss": 0.4065,
      "num_input_tokens_seen": 13648368,
      "step": 20875
    },
    {
      "epoch": 10.943396226415095,
      "grad_norm": 0.3787632882595062,
      "learning_rate": 2.5249263281389923e-05,
      "loss": 0.4923,
      "num_input_tokens_seen": 13650704,
      "step": 20880
    },
    {
      "epoch": 10.94601677148847,
      "grad_norm": 0.5211181640625,
      "learning_rate": 2.5237829539149004e-05,
      "loss": 0.5874,
      "num_input_tokens_seen": 13653552,
      "step": 20885
    },
    {
      "epoch": 10.948637316561845,
      "grad_norm": 0.30769553780555725,
      "learning_rate": 2.5226395747156866e-05,
      "loss": 0.4784,
      "num_input_tokens_seen": 13657488,
      "step": 20890
    },
    {
      "epoch": 10.95125786163522,
      "grad_norm": 0.583493173122406,
      "learning_rate": 2.5214961907805317e-05,
      "loss": 0.4384,
      "num_input_tokens_seen": 13660976,
      "step": 20895
    },
    {
      "epoch": 10.953878406708595,
      "grad_norm": 0.36636224389076233,
      "learning_rate": 2.520352802348619e-05,
      "loss": 0.3879,
      "num_input_tokens_seen": 13664080,
      "step": 20900
    },
    {
      "epoch": 10.95649895178197,
      "grad_norm": 0.49774348735809326,
      "learning_rate": 2.5192094096591323e-05,
      "loss": 0.6161,
      "num_input_tokens_seen": 13666960,
      "step": 20905
    },
    {
      "epoch": 10.959119496855346,
      "grad_norm": 0.29757022857666016,
      "learning_rate": 2.5180660129512556e-05,
      "loss": 0.4121,
      "num_input_tokens_seen": 13670832,
      "step": 20910
    },
    {
      "epoch": 10.96174004192872,
      "grad_norm": 0.7534888386726379,
      "learning_rate": 2.516922612464176e-05,
      "loss": 0.4368,
      "num_input_tokens_seen": 13673456,
      "step": 20915
    },
    {
      "epoch": 10.964360587002096,
      "grad_norm": 0.5603316426277161,
      "learning_rate": 2.5157792084370783e-05,
      "loss": 0.3825,
      "num_input_tokens_seen": 13676208,
      "step": 20920
    },
    {
      "epoch": 10.966981132075471,
      "grad_norm": 0.40992850065231323,
      "learning_rate": 2.5146358011091502e-05,
      "loss": 0.5235,
      "num_input_tokens_seen": 13679184,
      "step": 20925
    },
    {
      "epoch": 10.969601677148846,
      "grad_norm": 0.5660022497177124,
      "learning_rate": 2.5134923907195805e-05,
      "loss": 0.5029,
      "num_input_tokens_seen": 13682896,
      "step": 20930
    },
    {
      "epoch": 10.972222222222221,
      "grad_norm": 0.486766517162323,
      "learning_rate": 2.5123489775075558e-05,
      "loss": 0.5632,
      "num_input_tokens_seen": 13686064,
      "step": 20935
    },
    {
      "epoch": 10.974842767295598,
      "grad_norm": 0.2934097647666931,
      "learning_rate": 2.511205561712266e-05,
      "loss": 0.5366,
      "num_input_tokens_seen": 13689424,
      "step": 20940
    },
    {
      "epoch": 10.977463312368974,
      "grad_norm": 0.30868789553642273,
      "learning_rate": 2.5100621435729004e-05,
      "loss": 0.5674,
      "num_input_tokens_seen": 13692496,
      "step": 20945
    },
    {
      "epoch": 10.980083857442349,
      "grad_norm": 0.9678673148155212,
      "learning_rate": 2.5089187233286492e-05,
      "loss": 0.4792,
      "num_input_tokens_seen": 13695728,
      "step": 20950
    },
    {
      "epoch": 10.982704402515724,
      "grad_norm": 0.4727373719215393,
      "learning_rate": 2.5077753012187026e-05,
      "loss": 0.5008,
      "num_input_tokens_seen": 13698640,
      "step": 20955
    },
    {
      "epoch": 10.985324947589099,
      "grad_norm": 1.124734878540039,
      "learning_rate": 2.506631877482252e-05,
      "loss": 0.4949,
      "num_input_tokens_seen": 13702608,
      "step": 20960
    },
    {
      "epoch": 10.987945492662474,
      "grad_norm": 0.43458473682403564,
      "learning_rate": 2.505488452358488e-05,
      "loss": 0.5439,
      "num_input_tokens_seen": 13706352,
      "step": 20965
    },
    {
      "epoch": 10.99056603773585,
      "grad_norm": 0.4882420003414154,
      "learning_rate": 2.504345026086603e-05,
      "loss": 0.4052,
      "num_input_tokens_seen": 13711056,
      "step": 20970
    },
    {
      "epoch": 10.993186582809225,
      "grad_norm": 0.5702618956565857,
      "learning_rate": 2.5032015989057877e-05,
      "loss": 0.3861,
      "num_input_tokens_seen": 13714480,
      "step": 20975
    },
    {
      "epoch": 10.9958071278826,
      "grad_norm": 0.30623921751976013,
      "learning_rate": 2.5020581710552348e-05,
      "loss": 0.4122,
      "num_input_tokens_seen": 13717136,
      "step": 20980
    },
    {
      "epoch": 10.998427672955975,
      "grad_norm": 0.4279302954673767,
      "learning_rate": 2.5009147427741358e-05,
      "loss": 0.5687,
      "num_input_tokens_seen": 13720464,
      "step": 20985
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.47615760564804077,
      "eval_runtime": 9.7492,
      "eval_samples_per_second": 86.981,
      "eval_steps_per_second": 21.745,
      "num_input_tokens_seen": 13721824,
      "step": 20988
    },
    {
      "epoch": 11.00104821802935,
      "grad_norm": 0.4203367829322815,
      "learning_rate": 2.4997713143016822e-05,
      "loss": 0.3895,
      "num_input_tokens_seen": 13723072,
      "step": 20990
    },
    {
      "epoch": 11.003668763102725,
      "grad_norm": 0.4165584146976471,
      "learning_rate": 2.498627885877068e-05,
      "loss": 0.4054,
      "num_input_tokens_seen": 13726400,
      "step": 20995
    },
    {
      "epoch": 11.0062893081761,
      "grad_norm": 0.717910885810852,
      "learning_rate": 2.4974844577394846e-05,
      "loss": 0.3686,
      "num_input_tokens_seen": 13729280,
      "step": 21000
    },
    {
      "epoch": 11.008909853249476,
      "grad_norm": 0.4590103328227997,
      "learning_rate": 2.496341030128123e-05,
      "loss": 0.3842,
      "num_input_tokens_seen": 13732416,
      "step": 21005
    },
    {
      "epoch": 11.01153039832285,
      "grad_norm": 0.6540840864181519,
      "learning_rate": 2.495197603282177e-05,
      "loss": 0.5756,
      "num_input_tokens_seen": 13735232,
      "step": 21010
    },
    {
      "epoch": 11.014150943396226,
      "grad_norm": 0.6741893291473389,
      "learning_rate": 2.4940541774408368e-05,
      "loss": 0.4362,
      "num_input_tokens_seen": 13738656,
      "step": 21015
    },
    {
      "epoch": 11.016771488469601,
      "grad_norm": 0.30248668789863586,
      "learning_rate": 2.4929107528432953e-05,
      "loss": 0.3914,
      "num_input_tokens_seen": 13741568,
      "step": 21020
    },
    {
      "epoch": 11.019392033542976,
      "grad_norm": 0.3717598021030426,
      "learning_rate": 2.4917673297287446e-05,
      "loss": 0.3543,
      "num_input_tokens_seen": 13744192,
      "step": 21025
    },
    {
      "epoch": 11.022012578616351,
      "grad_norm": 2.852663993835449,
      "learning_rate": 2.4906239083363728e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 13746656,
      "step": 21030
    },
    {
      "epoch": 11.024633123689727,
      "grad_norm": 0.715964138507843,
      "learning_rate": 2.4894804889053732e-05,
      "loss": 0.4902,
      "num_input_tokens_seen": 13749920,
      "step": 21035
    },
    {
      "epoch": 11.027253668763104,
      "grad_norm": 0.4832131564617157,
      "learning_rate": 2.4883370716749347e-05,
      "loss": 0.5841,
      "num_input_tokens_seen": 13752864,
      "step": 21040
    },
    {
      "epoch": 11.029874213836479,
      "grad_norm": 0.6245779395103455,
      "learning_rate": 2.4871936568842486e-05,
      "loss": 0.4237,
      "num_input_tokens_seen": 13756192,
      "step": 21045
    },
    {
      "epoch": 11.032494758909854,
      "grad_norm": 0.5609173774719238,
      "learning_rate": 2.4860502447725033e-05,
      "loss": 0.4222,
      "num_input_tokens_seen": 13759968,
      "step": 21050
    },
    {
      "epoch": 11.035115303983229,
      "grad_norm": 0.3223415017127991,
      "learning_rate": 2.4849068355788873e-05,
      "loss": 0.4242,
      "num_input_tokens_seen": 13763072,
      "step": 21055
    },
    {
      "epoch": 11.037735849056604,
      "grad_norm": 0.5218358635902405,
      "learning_rate": 2.4837634295425892e-05,
      "loss": 0.3728,
      "num_input_tokens_seen": 13765728,
      "step": 21060
    },
    {
      "epoch": 11.04035639412998,
      "grad_norm": 0.37495341897010803,
      "learning_rate": 2.4826200269027957e-05,
      "loss": 0.4416,
      "num_input_tokens_seen": 13769504,
      "step": 21065
    },
    {
      "epoch": 11.042976939203355,
      "grad_norm": 0.42782461643218994,
      "learning_rate": 2.481476627898695e-05,
      "loss": 0.5025,
      "num_input_tokens_seen": 13773760,
      "step": 21070
    },
    {
      "epoch": 11.04559748427673,
      "grad_norm": 0.39718398451805115,
      "learning_rate": 2.4803332327694724e-05,
      "loss": 0.4427,
      "num_input_tokens_seen": 13777088,
      "step": 21075
    },
    {
      "epoch": 11.048218029350105,
      "grad_norm": 0.8185243010520935,
      "learning_rate": 2.4791898417543136e-05,
      "loss": 0.4308,
      "num_input_tokens_seen": 13781120,
      "step": 21080
    },
    {
      "epoch": 11.05083857442348,
      "grad_norm": 0.5432584881782532,
      "learning_rate": 2.4780464550924014e-05,
      "loss": 0.4343,
      "num_input_tokens_seen": 13784064,
      "step": 21085
    },
    {
      "epoch": 11.053459119496855,
      "grad_norm": 1.4067777395248413,
      "learning_rate": 2.4769030730229213e-05,
      "loss": 0.4956,
      "num_input_tokens_seen": 13787104,
      "step": 21090
    },
    {
      "epoch": 11.05607966457023,
      "grad_norm": 0.5101656317710876,
      "learning_rate": 2.475759695785054e-05,
      "loss": 0.4509,
      "num_input_tokens_seen": 13789824,
      "step": 21095
    },
    {
      "epoch": 11.058700209643606,
      "grad_norm": 0.4391002058982849,
      "learning_rate": 2.4746163236179822e-05,
      "loss": 0.5251,
      "num_input_tokens_seen": 13793152,
      "step": 21100
    },
    {
      "epoch": 11.06132075471698,
      "grad_norm": 0.3542288541793823,
      "learning_rate": 2.4734729567608858e-05,
      "loss": 0.5786,
      "num_input_tokens_seen": 13796288,
      "step": 21105
    },
    {
      "epoch": 11.063941299790356,
      "grad_norm": 0.4983022212982178,
      "learning_rate": 2.472329595452945e-05,
      "loss": 0.5256,
      "num_input_tokens_seen": 13799072,
      "step": 21110
    },
    {
      "epoch": 11.066561844863731,
      "grad_norm": 0.3404344916343689,
      "learning_rate": 2.4711862399333367e-05,
      "loss": 0.4142,
      "num_input_tokens_seen": 13802560,
      "step": 21115
    },
    {
      "epoch": 11.069182389937106,
      "grad_norm": 0.3215651214122772,
      "learning_rate": 2.4700428904412377e-05,
      "loss": 0.4435,
      "num_input_tokens_seen": 13806592,
      "step": 21120
    },
    {
      "epoch": 11.071802935010481,
      "grad_norm": 0.5021198391914368,
      "learning_rate": 2.4688995472158245e-05,
      "loss": 0.3755,
      "num_input_tokens_seen": 13809568,
      "step": 21125
    },
    {
      "epoch": 11.074423480083857,
      "grad_norm": 0.5345426797866821,
      "learning_rate": 2.4677562104962712e-05,
      "loss": 0.3595,
      "num_input_tokens_seen": 13812960,
      "step": 21130
    },
    {
      "epoch": 11.077044025157234,
      "grad_norm": 0.36736372113227844,
      "learning_rate": 2.4666128805217513e-05,
      "loss": 0.6643,
      "num_input_tokens_seen": 13815712,
      "step": 21135
    },
    {
      "epoch": 11.079664570230609,
      "grad_norm": 0.8648386597633362,
      "learning_rate": 2.465469557531436e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 13818560,
      "step": 21140
    },
    {
      "epoch": 11.082285115303984,
      "grad_norm": 0.3657662570476532,
      "learning_rate": 2.4643262417644944e-05,
      "loss": 0.3927,
      "num_input_tokens_seen": 13823264,
      "step": 21145
    },
    {
      "epoch": 11.084905660377359,
      "grad_norm": 0.3638342618942261,
      "learning_rate": 2.463182933460096e-05,
      "loss": 0.461,
      "num_input_tokens_seen": 13826464,
      "step": 21150
    },
    {
      "epoch": 11.087526205450734,
      "grad_norm": 0.5711155533790588,
      "learning_rate": 2.4620396328574087e-05,
      "loss": 0.3872,
      "num_input_tokens_seen": 13829280,
      "step": 21155
    },
    {
      "epoch": 11.09014675052411,
      "grad_norm": 0.6271826028823853,
      "learning_rate": 2.4608963401955965e-05,
      "loss": 0.3982,
      "num_input_tokens_seen": 13833344,
      "step": 21160
    },
    {
      "epoch": 11.092767295597485,
      "grad_norm": 0.21950778365135193,
      "learning_rate": 2.4597530557138247e-05,
      "loss": 0.4775,
      "num_input_tokens_seen": 13837088,
      "step": 21165
    },
    {
      "epoch": 11.09538784067086,
      "grad_norm": 0.6414963006973267,
      "learning_rate": 2.4586097796512538e-05,
      "loss": 0.5005,
      "num_input_tokens_seen": 13839488,
      "step": 21170
    },
    {
      "epoch": 11.098008385744235,
      "grad_norm": 0.41581907868385315,
      "learning_rate": 2.457466512247045e-05,
      "loss": 0.3597,
      "num_input_tokens_seen": 13842528,
      "step": 21175
    },
    {
      "epoch": 11.10062893081761,
      "grad_norm": 0.35666853189468384,
      "learning_rate": 2.4563232537403563e-05,
      "loss": 0.3322,
      "num_input_tokens_seen": 13845312,
      "step": 21180
    },
    {
      "epoch": 11.103249475890985,
      "grad_norm": 0.43919315934181213,
      "learning_rate": 2.4551800043703442e-05,
      "loss": 0.3868,
      "num_input_tokens_seen": 13848256,
      "step": 21185
    },
    {
      "epoch": 11.10587002096436,
      "grad_norm": 0.3260399103164673,
      "learning_rate": 2.454036764376165e-05,
      "loss": 0.4227,
      "num_input_tokens_seen": 13852064,
      "step": 21190
    },
    {
      "epoch": 11.108490566037736,
      "grad_norm": 0.4416593909263611,
      "learning_rate": 2.4528935339969705e-05,
      "loss": 0.5192,
      "num_input_tokens_seen": 13855040,
      "step": 21195
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 0.30428099632263184,
      "learning_rate": 2.45175031347191e-05,
      "loss": 0.4912,
      "num_input_tokens_seen": 13858528,
      "step": 21200
    },
    {
      "epoch": 11.113731656184486,
      "grad_norm": 0.4829815924167633,
      "learning_rate": 2.4506071030401344e-05,
      "loss": 0.4631,
      "num_input_tokens_seen": 13862048,
      "step": 21205
    },
    {
      "epoch": 11.116352201257861,
      "grad_norm": 0.48058539628982544,
      "learning_rate": 2.4494639029407884e-05,
      "loss": 0.4254,
      "num_input_tokens_seen": 13864864,
      "step": 21210
    },
    {
      "epoch": 11.118972746331236,
      "grad_norm": 0.31096991896629333,
      "learning_rate": 2.448320713413018e-05,
      "loss": 0.3887,
      "num_input_tokens_seen": 13868256,
      "step": 21215
    },
    {
      "epoch": 11.121593291404611,
      "grad_norm": 0.8774513006210327,
      "learning_rate": 2.4471775346959654e-05,
      "loss": 0.4644,
      "num_input_tokens_seen": 13870944,
      "step": 21220
    },
    {
      "epoch": 11.124213836477987,
      "grad_norm": 0.35695576667785645,
      "learning_rate": 2.446034367028769e-05,
      "loss": 0.4263,
      "num_input_tokens_seen": 13874016,
      "step": 21225
    },
    {
      "epoch": 11.126834381551364,
      "grad_norm": 0.5526747107505798,
      "learning_rate": 2.444891210650567e-05,
      "loss": 0.556,
      "num_input_tokens_seen": 13881184,
      "step": 21230
    },
    {
      "epoch": 11.129454926624739,
      "grad_norm": 0.3839958906173706,
      "learning_rate": 2.4437480658004958e-05,
      "loss": 0.4971,
      "num_input_tokens_seen": 13884224,
      "step": 21235
    },
    {
      "epoch": 11.132075471698114,
      "grad_norm": 0.43727660179138184,
      "learning_rate": 2.4426049327176868e-05,
      "loss": 0.3494,
      "num_input_tokens_seen": 13888064,
      "step": 21240
    },
    {
      "epoch": 11.134696016771489,
      "grad_norm": 0.41918131709098816,
      "learning_rate": 2.4414618116412716e-05,
      "loss": 0.3352,
      "num_input_tokens_seen": 13890912,
      "step": 21245
    },
    {
      "epoch": 11.137316561844864,
      "grad_norm": 0.47745421528816223,
      "learning_rate": 2.440318702810378e-05,
      "loss": 0.5248,
      "num_input_tokens_seen": 13894432,
      "step": 21250
    },
    {
      "epoch": 11.13993710691824,
      "grad_norm": 0.6148984432220459,
      "learning_rate": 2.43917560646413e-05,
      "loss": 0.5292,
      "num_input_tokens_seen": 13898048,
      "step": 21255
    },
    {
      "epoch": 11.142557651991615,
      "grad_norm": 1.0454871654510498,
      "learning_rate": 2.4380325228416518e-05,
      "loss": 0.6483,
      "num_input_tokens_seen": 13901184,
      "step": 21260
    },
    {
      "epoch": 11.14517819706499,
      "grad_norm": 0.4975892901420593,
      "learning_rate": 2.4368894521820624e-05,
      "loss": 0.3118,
      "num_input_tokens_seen": 13904864,
      "step": 21265
    },
    {
      "epoch": 11.147798742138365,
      "grad_norm": 0.42321741580963135,
      "learning_rate": 2.43574639472448e-05,
      "loss": 0.3658,
      "num_input_tokens_seen": 13908096,
      "step": 21270
    },
    {
      "epoch": 11.15041928721174,
      "grad_norm": 0.3174971640110016,
      "learning_rate": 2.4346033507080178e-05,
      "loss": 0.3149,
      "num_input_tokens_seen": 13911392,
      "step": 21275
    },
    {
      "epoch": 11.153039832285115,
      "grad_norm": 0.3643586039543152,
      "learning_rate": 2.4334603203717903e-05,
      "loss": 0.5221,
      "num_input_tokens_seen": 13915104,
      "step": 21280
    },
    {
      "epoch": 11.15566037735849,
      "grad_norm": 0.4143342077732086,
      "learning_rate": 2.432317303954904e-05,
      "loss": 0.4435,
      "num_input_tokens_seen": 13918176,
      "step": 21285
    },
    {
      "epoch": 11.158280922431866,
      "grad_norm": 0.32011643052101135,
      "learning_rate": 2.4311743016964648e-05,
      "loss": 0.363,
      "num_input_tokens_seen": 13921120,
      "step": 21290
    },
    {
      "epoch": 11.16090146750524,
      "grad_norm": 0.42866817116737366,
      "learning_rate": 2.4300313138355772e-05,
      "loss": 0.3731,
      "num_input_tokens_seen": 13924224,
      "step": 21295
    },
    {
      "epoch": 11.163522012578616,
      "grad_norm": 0.3653757870197296,
      "learning_rate": 2.4288883406113396e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 13927488,
      "step": 21300
    },
    {
      "epoch": 11.166142557651991,
      "grad_norm": 0.6781365871429443,
      "learning_rate": 2.4277453822628512e-05,
      "loss": 0.3466,
      "num_input_tokens_seen": 13930880,
      "step": 21305
    },
    {
      "epoch": 11.168763102725366,
      "grad_norm": 0.4538199007511139,
      "learning_rate": 2.4266024390292037e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 13933728,
      "step": 21310
    },
    {
      "epoch": 11.171383647798741,
      "grad_norm": 0.5028011202812195,
      "learning_rate": 2.425459511149489e-05,
      "loss": 0.492,
      "num_input_tokens_seen": 13937664,
      "step": 21315
    },
    {
      "epoch": 11.174004192872117,
      "grad_norm": 0.27711716294288635,
      "learning_rate": 2.424316598862793e-05,
      "loss": 0.3609,
      "num_input_tokens_seen": 13940544,
      "step": 21320
    },
    {
      "epoch": 11.176624737945493,
      "grad_norm": 0.284378319978714,
      "learning_rate": 2.4231737024082023e-05,
      "loss": 0.427,
      "num_input_tokens_seen": 13943712,
      "step": 21325
    },
    {
      "epoch": 11.179245283018869,
      "grad_norm": 0.34716805815696716,
      "learning_rate": 2.4220308220247954e-05,
      "loss": 0.3644,
      "num_input_tokens_seen": 13947200,
      "step": 21330
    },
    {
      "epoch": 11.181865828092244,
      "grad_norm": 0.7437220811843872,
      "learning_rate": 2.4208879579516528e-05,
      "loss": 0.803,
      "num_input_tokens_seen": 13950240,
      "step": 21335
    },
    {
      "epoch": 11.184486373165619,
      "grad_norm": 0.3293205499649048,
      "learning_rate": 2.419745110427846e-05,
      "loss": 0.3388,
      "num_input_tokens_seen": 13953440,
      "step": 21340
    },
    {
      "epoch": 11.187106918238994,
      "grad_norm": 0.35926979780197144,
      "learning_rate": 2.418602279692446e-05,
      "loss": 0.4088,
      "num_input_tokens_seen": 13956416,
      "step": 21345
    },
    {
      "epoch": 11.18972746331237,
      "grad_norm": 0.34468886256217957,
      "learning_rate": 2.417459465984521e-05,
      "loss": 0.415,
      "num_input_tokens_seen": 13959840,
      "step": 21350
    },
    {
      "epoch": 11.192348008385745,
      "grad_norm": 0.7806999087333679,
      "learning_rate": 2.4163166695431336e-05,
      "loss": 0.5928,
      "num_input_tokens_seen": 13963168,
      "step": 21355
    },
    {
      "epoch": 11.19496855345912,
      "grad_norm": 0.44203025102615356,
      "learning_rate": 2.4151738906073446e-05,
      "loss": 0.3419,
      "num_input_tokens_seen": 13965632,
      "step": 21360
    },
    {
      "epoch": 11.197589098532495,
      "grad_norm": 0.467512309551239,
      "learning_rate": 2.4140311294162105e-05,
      "loss": 0.4453,
      "num_input_tokens_seen": 13969184,
      "step": 21365
    },
    {
      "epoch": 11.20020964360587,
      "grad_norm": 0.4233495593070984,
      "learning_rate": 2.412888386208782e-05,
      "loss": 0.4703,
      "num_input_tokens_seen": 13972800,
      "step": 21370
    },
    {
      "epoch": 11.202830188679245,
      "grad_norm": 0.9369889497756958,
      "learning_rate": 2.4117456612241103e-05,
      "loss": 0.6175,
      "num_input_tokens_seen": 13976128,
      "step": 21375
    },
    {
      "epoch": 11.20545073375262,
      "grad_norm": 0.2513447701931,
      "learning_rate": 2.410602954701239e-05,
      "loss": 0.448,
      "num_input_tokens_seen": 13979424,
      "step": 21380
    },
    {
      "epoch": 11.208071278825996,
      "grad_norm": 0.4394517242908478,
      "learning_rate": 2.409460266879209e-05,
      "loss": 0.3467,
      "num_input_tokens_seen": 13981856,
      "step": 21385
    },
    {
      "epoch": 11.21069182389937,
      "grad_norm": 0.40184980630874634,
      "learning_rate": 2.4083175979970592e-05,
      "loss": 0.3363,
      "num_input_tokens_seen": 13985248,
      "step": 21390
    },
    {
      "epoch": 11.213312368972746,
      "grad_norm": 0.3409738540649414,
      "learning_rate": 2.4071749482938226e-05,
      "loss": 0.4859,
      "num_input_tokens_seen": 13988576,
      "step": 21395
    },
    {
      "epoch": 11.215932914046121,
      "grad_norm": 0.8358817100524902,
      "learning_rate": 2.4060323180085266e-05,
      "loss": 0.4908,
      "num_input_tokens_seen": 13991232,
      "step": 21400
    },
    {
      "epoch": 11.218553459119496,
      "grad_norm": 0.45925310254096985,
      "learning_rate": 2.404889707380198e-05,
      "loss": 0.4234,
      "num_input_tokens_seen": 13994592,
      "step": 21405
    },
    {
      "epoch": 11.221174004192871,
      "grad_norm": 0.7932262420654297,
      "learning_rate": 2.403747116647857e-05,
      "loss": 0.414,
      "num_input_tokens_seen": 13997280,
      "step": 21410
    },
    {
      "epoch": 11.223794549266247,
      "grad_norm": 0.4961133897304535,
      "learning_rate": 2.4026045460505216e-05,
      "loss": 0.4368,
      "num_input_tokens_seen": 14000384,
      "step": 21415
    },
    {
      "epoch": 11.226415094339623,
      "grad_norm": 0.41194209456443787,
      "learning_rate": 2.4014619958272047e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 14003872,
      "step": 21420
    },
    {
      "epoch": 11.229035639412999,
      "grad_norm": 0.392871618270874,
      "learning_rate": 2.4003194662169132e-05,
      "loss": 0.3949,
      "num_input_tokens_seen": 14006976,
      "step": 21425
    },
    {
      "epoch": 11.231656184486374,
      "grad_norm": 0.39715781807899475,
      "learning_rate": 2.3991769574586522e-05,
      "loss": 0.73,
      "num_input_tokens_seen": 14009472,
      "step": 21430
    },
    {
      "epoch": 11.234276729559749,
      "grad_norm": 0.7185111045837402,
      "learning_rate": 2.3980344697914215e-05,
      "loss": 0.4049,
      "num_input_tokens_seen": 14012480,
      "step": 21435
    },
    {
      "epoch": 11.236897274633124,
      "grad_norm": 0.36953142285346985,
      "learning_rate": 2.3968920034542168e-05,
      "loss": 0.3568,
      "num_input_tokens_seen": 14015072,
      "step": 21440
    },
    {
      "epoch": 11.2395178197065,
      "grad_norm": 0.502988874912262,
      "learning_rate": 2.3957495586860278e-05,
      "loss": 0.4643,
      "num_input_tokens_seen": 14018496,
      "step": 21445
    },
    {
      "epoch": 11.242138364779874,
      "grad_norm": 0.5993635058403015,
      "learning_rate": 2.394607135725843e-05,
      "loss": 0.5334,
      "num_input_tokens_seen": 14021600,
      "step": 21450
    },
    {
      "epoch": 11.24475890985325,
      "grad_norm": 0.4532373547554016,
      "learning_rate": 2.3934647348126426e-05,
      "loss": 0.4558,
      "num_input_tokens_seen": 14024544,
      "step": 21455
    },
    {
      "epoch": 11.247379454926625,
      "grad_norm": 0.43888503313064575,
      "learning_rate": 2.3923223561854035e-05,
      "loss": 0.4445,
      "num_input_tokens_seen": 14027968,
      "step": 21460
    },
    {
      "epoch": 11.25,
      "grad_norm": 0.5838133096694946,
      "learning_rate": 2.391180000083099e-05,
      "loss": 0.4834,
      "num_input_tokens_seen": 14030688,
      "step": 21465
    },
    {
      "epoch": 11.252620545073375,
      "grad_norm": 0.379025936126709,
      "learning_rate": 2.390037666744698e-05,
      "loss": 0.4068,
      "num_input_tokens_seen": 14033408,
      "step": 21470
    },
    {
      "epoch": 11.25524109014675,
      "grad_norm": 0.3597854673862457,
      "learning_rate": 2.388895356409162e-05,
      "loss": 0.4021,
      "num_input_tokens_seen": 14036416,
      "step": 21475
    },
    {
      "epoch": 11.257861635220126,
      "grad_norm": 0.48029541969299316,
      "learning_rate": 2.3877530693154506e-05,
      "loss": 0.4262,
      "num_input_tokens_seen": 14039872,
      "step": 21480
    },
    {
      "epoch": 11.2604821802935,
      "grad_norm": 0.3932454288005829,
      "learning_rate": 2.3866108057025162e-05,
      "loss": 0.4692,
      "num_input_tokens_seen": 14042624,
      "step": 21485
    },
    {
      "epoch": 11.263102725366876,
      "grad_norm": 0.7005370855331421,
      "learning_rate": 2.385468565809307e-05,
      "loss": 0.4916,
      "num_input_tokens_seen": 14045984,
      "step": 21490
    },
    {
      "epoch": 11.265723270440251,
      "grad_norm": 0.40527018904685974,
      "learning_rate": 2.3843263498747675e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 14049152,
      "step": 21495
    },
    {
      "epoch": 11.268343815513626,
      "grad_norm": 0.5369002223014832,
      "learning_rate": 2.383184158137836e-05,
      "loss": 0.4291,
      "num_input_tokens_seen": 14052320,
      "step": 21500
    },
    {
      "epoch": 11.270964360587001,
      "grad_norm": 0.33932629227638245,
      "learning_rate": 2.3820419908374465e-05,
      "loss": 0.4321,
      "num_input_tokens_seen": 14055584,
      "step": 21505
    },
    {
      "epoch": 11.273584905660377,
      "grad_norm": 0.5996184945106506,
      "learning_rate": 2.3808998482125265e-05,
      "loss": 0.5423,
      "num_input_tokens_seen": 14058816,
      "step": 21510
    },
    {
      "epoch": 11.276205450733753,
      "grad_norm": 0.3403533399105072,
      "learning_rate": 2.3797577305019987e-05,
      "loss": 0.6222,
      "num_input_tokens_seen": 14062528,
      "step": 21515
    },
    {
      "epoch": 11.278825995807129,
      "grad_norm": 0.7885410189628601,
      "learning_rate": 2.378615637944783e-05,
      "loss": 0.603,
      "num_input_tokens_seen": 14065280,
      "step": 21520
    },
    {
      "epoch": 11.281446540880504,
      "grad_norm": 0.8270146250724792,
      "learning_rate": 2.3774735707797895e-05,
      "loss": 0.411,
      "num_input_tokens_seen": 14068160,
      "step": 21525
    },
    {
      "epoch": 11.284067085953879,
      "grad_norm": 0.4686325788497925,
      "learning_rate": 2.376331529245928e-05,
      "loss": 0.4423,
      "num_input_tokens_seen": 14071712,
      "step": 21530
    },
    {
      "epoch": 11.286687631027254,
      "grad_norm": 0.29727423191070557,
      "learning_rate": 2.3751895135821006e-05,
      "loss": 0.3806,
      "num_input_tokens_seen": 14074912,
      "step": 21535
    },
    {
      "epoch": 11.28930817610063,
      "grad_norm": 0.643600583076477,
      "learning_rate": 2.374047524027202e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 14078912,
      "step": 21540
    },
    {
      "epoch": 11.291928721174004,
      "grad_norm": 0.3527523875236511,
      "learning_rate": 2.3729055608201243e-05,
      "loss": 0.3982,
      "num_input_tokens_seen": 14082304,
      "step": 21545
    },
    {
      "epoch": 11.29454926624738,
      "grad_norm": 0.5557048320770264,
      "learning_rate": 2.3717636241997537e-05,
      "loss": 0.4524,
      "num_input_tokens_seen": 14085056,
      "step": 21550
    },
    {
      "epoch": 11.297169811320755,
      "grad_norm": 0.3733016848564148,
      "learning_rate": 2.3706217144049696e-05,
      "loss": 0.504,
      "num_input_tokens_seen": 14088128,
      "step": 21555
    },
    {
      "epoch": 11.29979035639413,
      "grad_norm": 0.3804794251918793,
      "learning_rate": 2.369479831674647e-05,
      "loss": 0.4026,
      "num_input_tokens_seen": 14090912,
      "step": 21560
    },
    {
      "epoch": 11.302410901467505,
      "grad_norm": 0.29118674993515015,
      "learning_rate": 2.3683379762476555e-05,
      "loss": 0.5797,
      "num_input_tokens_seen": 14094464,
      "step": 21565
    },
    {
      "epoch": 11.30503144654088,
      "grad_norm": 0.5095493197441101,
      "learning_rate": 2.3671961483628562e-05,
      "loss": 0.4519,
      "num_input_tokens_seen": 14097280,
      "step": 21570
    },
    {
      "epoch": 11.307651991614255,
      "grad_norm": 0.4233874976634979,
      "learning_rate": 2.3660543482591078e-05,
      "loss": 0.4821,
      "num_input_tokens_seen": 14100224,
      "step": 21575
    },
    {
      "epoch": 11.31027253668763,
      "grad_norm": 0.29022321105003357,
      "learning_rate": 2.364912576175261e-05,
      "loss": 0.4323,
      "num_input_tokens_seen": 14103744,
      "step": 21580
    },
    {
      "epoch": 11.312893081761006,
      "grad_norm": 0.33656856417655945,
      "learning_rate": 2.363770832350163e-05,
      "loss": 0.3675,
      "num_input_tokens_seen": 14107008,
      "step": 21585
    },
    {
      "epoch": 11.315513626834381,
      "grad_norm": 0.42276811599731445,
      "learning_rate": 2.362629117022653e-05,
      "loss": 0.417,
      "num_input_tokens_seen": 14110176,
      "step": 21590
    },
    {
      "epoch": 11.318134171907756,
      "grad_norm": 0.5372228622436523,
      "learning_rate": 2.3614874304315633e-05,
      "loss": 0.6346,
      "num_input_tokens_seen": 14115968,
      "step": 21595
    },
    {
      "epoch": 11.320754716981131,
      "grad_norm": 0.24953535199165344,
      "learning_rate": 2.3603457728157237e-05,
      "loss": 0.3685,
      "num_input_tokens_seen": 14120864,
      "step": 21600
    },
    {
      "epoch": 11.323375262054507,
      "grad_norm": 0.3224717080593109,
      "learning_rate": 2.3592041444139547e-05,
      "loss": 0.5423,
      "num_input_tokens_seen": 14124576,
      "step": 21605
    },
    {
      "epoch": 11.325995807127883,
      "grad_norm": 0.32072713971138,
      "learning_rate": 2.358062545465073e-05,
      "loss": 0.4971,
      "num_input_tokens_seen": 14127648,
      "step": 21610
    },
    {
      "epoch": 11.328616352201259,
      "grad_norm": 0.4368753731250763,
      "learning_rate": 2.3569209762078866e-05,
      "loss": 0.4049,
      "num_input_tokens_seen": 14130400,
      "step": 21615
    },
    {
      "epoch": 11.331236897274634,
      "grad_norm": 0.2884124219417572,
      "learning_rate": 2.355779436881201e-05,
      "loss": 0.4209,
      "num_input_tokens_seen": 14134080,
      "step": 21620
    },
    {
      "epoch": 11.333857442348009,
      "grad_norm": 0.5427463054656982,
      "learning_rate": 2.3546379277238107e-05,
      "loss": 0.3987,
      "num_input_tokens_seen": 14138080,
      "step": 21625
    },
    {
      "epoch": 11.336477987421384,
      "grad_norm": 0.33838382363319397,
      "learning_rate": 2.353496448974508e-05,
      "loss": 0.3563,
      "num_input_tokens_seen": 14142432,
      "step": 21630
    },
    {
      "epoch": 11.33909853249476,
      "grad_norm": 0.6228436827659607,
      "learning_rate": 2.3523550008720763e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 14145440,
      "step": 21635
    },
    {
      "epoch": 11.341719077568134,
      "grad_norm": 0.43320029973983765,
      "learning_rate": 2.3512135836552945e-05,
      "loss": 0.4285,
      "num_input_tokens_seen": 14148672,
      "step": 21640
    },
    {
      "epoch": 11.34433962264151,
      "grad_norm": 0.3169589340686798,
      "learning_rate": 2.350072197562933e-05,
      "loss": 0.361,
      "num_input_tokens_seen": 14152000,
      "step": 21645
    },
    {
      "epoch": 11.346960167714885,
      "grad_norm": 0.44323259592056274,
      "learning_rate": 2.3489308428337586e-05,
      "loss": 0.5015,
      "num_input_tokens_seen": 14154912,
      "step": 21650
    },
    {
      "epoch": 11.34958071278826,
      "grad_norm": 0.3824898600578308,
      "learning_rate": 2.3477895197065275e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 14158208,
      "step": 21655
    },
    {
      "epoch": 11.352201257861635,
      "grad_norm": 0.49806904792785645,
      "learning_rate": 2.346648228419992e-05,
      "loss": 0.4829,
      "num_input_tokens_seen": 14161920,
      "step": 21660
    },
    {
      "epoch": 11.35482180293501,
      "grad_norm": 0.544448733329773,
      "learning_rate": 2.3455069692128982e-05,
      "loss": 0.4026,
      "num_input_tokens_seen": 14165120,
      "step": 21665
    },
    {
      "epoch": 11.357442348008385,
      "grad_norm": 0.43867599964141846,
      "learning_rate": 2.3443657423239832e-05,
      "loss": 0.4712,
      "num_input_tokens_seen": 14168000,
      "step": 21670
    },
    {
      "epoch": 11.36006289308176,
      "grad_norm": 0.35390758514404297,
      "learning_rate": 2.34322454799198e-05,
      "loss": 0.3661,
      "num_input_tokens_seen": 14171584,
      "step": 21675
    },
    {
      "epoch": 11.362683438155136,
      "grad_norm": 0.6419451832771301,
      "learning_rate": 2.3420833864556133e-05,
      "loss": 0.4456,
      "num_input_tokens_seen": 14174336,
      "step": 21680
    },
    {
      "epoch": 11.365303983228511,
      "grad_norm": 0.5816247463226318,
      "learning_rate": 2.3409422579535995e-05,
      "loss": 0.4587,
      "num_input_tokens_seen": 14177344,
      "step": 21685
    },
    {
      "epoch": 11.367924528301886,
      "grad_norm": 0.4382965564727783,
      "learning_rate": 2.339801162724651e-05,
      "loss": 0.4782,
      "num_input_tokens_seen": 14180608,
      "step": 21690
    },
    {
      "epoch": 11.370545073375261,
      "grad_norm": 0.3443540036678314,
      "learning_rate": 2.338660101007471e-05,
      "loss": 0.5005,
      "num_input_tokens_seen": 14183872,
      "step": 21695
    },
    {
      "epoch": 11.373165618448636,
      "grad_norm": 0.595944344997406,
      "learning_rate": 2.3375190730407578e-05,
      "loss": 0.4947,
      "num_input_tokens_seen": 14186720,
      "step": 21700
    },
    {
      "epoch": 11.375786163522013,
      "grad_norm": 0.39689934253692627,
      "learning_rate": 2.336378079063202e-05,
      "loss": 0.5496,
      "num_input_tokens_seen": 14190336,
      "step": 21705
    },
    {
      "epoch": 11.378406708595389,
      "grad_norm": 0.39369404315948486,
      "learning_rate": 2.335237119313484e-05,
      "loss": 0.4479,
      "num_input_tokens_seen": 14193760,
      "step": 21710
    },
    {
      "epoch": 11.381027253668764,
      "grad_norm": 0.6301022171974182,
      "learning_rate": 2.3340961940302807e-05,
      "loss": 0.4478,
      "num_input_tokens_seen": 14196096,
      "step": 21715
    },
    {
      "epoch": 11.383647798742139,
      "grad_norm": 0.6095010638237,
      "learning_rate": 2.3329553034522616e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 14199296,
      "step": 21720
    },
    {
      "epoch": 11.386268343815514,
      "grad_norm": 0.32280781865119934,
      "learning_rate": 2.3318144478180866e-05,
      "loss": 0.442,
      "num_input_tokens_seen": 14202528,
      "step": 21725
    },
    {
      "epoch": 11.38888888888889,
      "grad_norm": 0.2881181538105011,
      "learning_rate": 2.330673627366411e-05,
      "loss": 0.465,
      "num_input_tokens_seen": 14205568,
      "step": 21730
    },
    {
      "epoch": 11.391509433962264,
      "grad_norm": 0.2792731821537018,
      "learning_rate": 2.329532842335881e-05,
      "loss": 0.4309,
      "num_input_tokens_seen": 14209056,
      "step": 21735
    },
    {
      "epoch": 11.39412997903564,
      "grad_norm": 0.42050495743751526,
      "learning_rate": 2.328392092965135e-05,
      "loss": 0.2314,
      "num_input_tokens_seen": 14214720,
      "step": 21740
    },
    {
      "epoch": 11.396750524109015,
      "grad_norm": 0.30393436551094055,
      "learning_rate": 2.3272513794928054e-05,
      "loss": 0.4426,
      "num_input_tokens_seen": 14217664,
      "step": 21745
    },
    {
      "epoch": 11.39937106918239,
      "grad_norm": 0.38278234004974365,
      "learning_rate": 2.326110702157516e-05,
      "loss": 0.4725,
      "num_input_tokens_seen": 14221216,
      "step": 21750
    },
    {
      "epoch": 11.401991614255765,
      "grad_norm": 0.6838162541389465,
      "learning_rate": 2.3249700611978847e-05,
      "loss": 0.4469,
      "num_input_tokens_seen": 14224000,
      "step": 21755
    },
    {
      "epoch": 11.40461215932914,
      "grad_norm": 0.3151335120201111,
      "learning_rate": 2.323829456852519e-05,
      "loss": 0.4398,
      "num_input_tokens_seen": 14227872,
      "step": 21760
    },
    {
      "epoch": 11.407232704402515,
      "grad_norm": 0.5228821039199829,
      "learning_rate": 2.3226888893600223e-05,
      "loss": 0.5195,
      "num_input_tokens_seen": 14231936,
      "step": 21765
    },
    {
      "epoch": 11.40985324947589,
      "grad_norm": 0.4474312663078308,
      "learning_rate": 2.3215483589589865e-05,
      "loss": 0.3397,
      "num_input_tokens_seen": 14235328,
      "step": 21770
    },
    {
      "epoch": 11.412473794549266,
      "grad_norm": 0.7928164601325989,
      "learning_rate": 2.3204078658879974e-05,
      "loss": 0.4292,
      "num_input_tokens_seen": 14238848,
      "step": 21775
    },
    {
      "epoch": 11.415094339622641,
      "grad_norm": 0.5910794138908386,
      "learning_rate": 2.3192674103856337e-05,
      "loss": 0.5571,
      "num_input_tokens_seen": 14242336,
      "step": 21780
    },
    {
      "epoch": 11.417714884696016,
      "grad_norm": 0.6662541031837463,
      "learning_rate": 2.318126992690467e-05,
      "loss": 0.463,
      "num_input_tokens_seen": 14245728,
      "step": 21785
    },
    {
      "epoch": 11.420335429769391,
      "grad_norm": 0.6107650995254517,
      "learning_rate": 2.3169866130410585e-05,
      "loss": 0.3155,
      "num_input_tokens_seen": 14248448,
      "step": 21790
    },
    {
      "epoch": 11.422955974842766,
      "grad_norm": 0.2659282386302948,
      "learning_rate": 2.315846271675962e-05,
      "loss": 0.4256,
      "num_input_tokens_seen": 14251648,
      "step": 21795
    },
    {
      "epoch": 11.425576519916142,
      "grad_norm": 0.33879348635673523,
      "learning_rate": 2.3147059688337244e-05,
      "loss": 0.3451,
      "num_input_tokens_seen": 14254624,
      "step": 21800
    },
    {
      "epoch": 11.428197064989519,
      "grad_norm": 0.6437053084373474,
      "learning_rate": 2.313565704752884e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 14257728,
      "step": 21805
    },
    {
      "epoch": 11.430817610062894,
      "grad_norm": 0.38323432207107544,
      "learning_rate": 2.3124254796719718e-05,
      "loss": 0.5105,
      "num_input_tokens_seen": 14262176,
      "step": 21810
    },
    {
      "epoch": 11.433438155136269,
      "grad_norm": 0.6220484972000122,
      "learning_rate": 2.3112852938295088e-05,
      "loss": 0.4385,
      "num_input_tokens_seen": 14265760,
      "step": 21815
    },
    {
      "epoch": 11.436058700209644,
      "grad_norm": 0.7790257334709167,
      "learning_rate": 2.3101451474640104e-05,
      "loss": 0.4994,
      "num_input_tokens_seen": 14268768,
      "step": 21820
    },
    {
      "epoch": 11.43867924528302,
      "grad_norm": 0.4776008129119873,
      "learning_rate": 2.3090050408139808e-05,
      "loss": 0.554,
      "num_input_tokens_seen": 14272064,
      "step": 21825
    },
    {
      "epoch": 11.441299790356394,
      "grad_norm": 0.40355297923088074,
      "learning_rate": 2.307864974117917e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 14275456,
      "step": 21830
    },
    {
      "epoch": 11.44392033542977,
      "grad_norm": 0.46282723546028137,
      "learning_rate": 2.3067249476143087e-05,
      "loss": 0.4998,
      "num_input_tokens_seen": 14279712,
      "step": 21835
    },
    {
      "epoch": 11.446540880503145,
      "grad_norm": 0.5235423445701599,
      "learning_rate": 2.3055849615416363e-05,
      "loss": 0.4279,
      "num_input_tokens_seen": 14283616,
      "step": 21840
    },
    {
      "epoch": 11.44916142557652,
      "grad_norm": 0.38045862317085266,
      "learning_rate": 2.304445016138372e-05,
      "loss": 0.3121,
      "num_input_tokens_seen": 14286720,
      "step": 21845
    },
    {
      "epoch": 11.451781970649895,
      "grad_norm": 0.8123760223388672,
      "learning_rate": 2.303305111642981e-05,
      "loss": 0.441,
      "num_input_tokens_seen": 14291232,
      "step": 21850
    },
    {
      "epoch": 11.45440251572327,
      "grad_norm": 0.7188053131103516,
      "learning_rate": 2.3021652482939144e-05,
      "loss": 0.4559,
      "num_input_tokens_seen": 14294496,
      "step": 21855
    },
    {
      "epoch": 11.457023060796645,
      "grad_norm": 0.27514058351516724,
      "learning_rate": 2.3010254263296214e-05,
      "loss": 0.4539,
      "num_input_tokens_seen": 14297088,
      "step": 21860
    },
    {
      "epoch": 11.45964360587002,
      "grad_norm": 0.45981380343437195,
      "learning_rate": 2.2998856459885396e-05,
      "loss": 0.3704,
      "num_input_tokens_seen": 14301056,
      "step": 21865
    },
    {
      "epoch": 11.462264150943396,
      "grad_norm": 0.4423467814922333,
      "learning_rate": 2.298745907509097e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 14303968,
      "step": 21870
    },
    {
      "epoch": 11.464884696016771,
      "grad_norm": 0.2298405021429062,
      "learning_rate": 2.2976062111297156e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 14308256,
      "step": 21875
    },
    {
      "epoch": 11.467505241090146,
      "grad_norm": 0.4139261841773987,
      "learning_rate": 2.296466557088805e-05,
      "loss": 0.5639,
      "num_input_tokens_seen": 14311424,
      "step": 21880
    },
    {
      "epoch": 11.470125786163521,
      "grad_norm": 0.43458813428878784,
      "learning_rate": 2.2953269456247688e-05,
      "loss": 0.4781,
      "num_input_tokens_seen": 14314752,
      "step": 21885
    },
    {
      "epoch": 11.472746331236896,
      "grad_norm": 0.3458916246891022,
      "learning_rate": 2.2941873769760006e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 14317760,
      "step": 21890
    },
    {
      "epoch": 11.475366876310272,
      "grad_norm": 0.587119996547699,
      "learning_rate": 2.2930478513808844e-05,
      "loss": 0.5754,
      "num_input_tokens_seen": 14321440,
      "step": 21895
    },
    {
      "epoch": 11.477987421383649,
      "grad_norm": 0.6964570879936218,
      "learning_rate": 2.2919083690777976e-05,
      "loss": 0.5281,
      "num_input_tokens_seen": 14324512,
      "step": 21900
    },
    {
      "epoch": 11.480607966457024,
      "grad_norm": 0.43953433632850647,
      "learning_rate": 2.2907689303051062e-05,
      "loss": 0.53,
      "num_input_tokens_seen": 14327648,
      "step": 21905
    },
    {
      "epoch": 11.483228511530399,
      "grad_norm": 0.38634830713272095,
      "learning_rate": 2.289629535301167e-05,
      "loss": 0.5732,
      "num_input_tokens_seen": 14331040,
      "step": 21910
    },
    {
      "epoch": 11.485849056603774,
      "grad_norm": 0.5712026357650757,
      "learning_rate": 2.288490184304329e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 14333920,
      "step": 21915
    },
    {
      "epoch": 11.48846960167715,
      "grad_norm": 0.6283664107322693,
      "learning_rate": 2.2873508775529308e-05,
      "loss": 0.4756,
      "num_input_tokens_seen": 14337056,
      "step": 21920
    },
    {
      "epoch": 11.491090146750524,
      "grad_norm": 0.436448335647583,
      "learning_rate": 2.2862116152853034e-05,
      "loss": 0.3283,
      "num_input_tokens_seen": 14339584,
      "step": 21925
    },
    {
      "epoch": 11.4937106918239,
      "grad_norm": 0.27539703249931335,
      "learning_rate": 2.285072397739767e-05,
      "loss": 0.3569,
      "num_input_tokens_seen": 14345024,
      "step": 21930
    },
    {
      "epoch": 11.496331236897275,
      "grad_norm": 0.34909385442733765,
      "learning_rate": 2.283933225154634e-05,
      "loss": 0.423,
      "num_input_tokens_seen": 14348288,
      "step": 21935
    },
    {
      "epoch": 11.49895178197065,
      "grad_norm": 1.197078824043274,
      "learning_rate": 2.2827940977682045e-05,
      "loss": 0.5919,
      "num_input_tokens_seen": 14352000,
      "step": 21940
    },
    {
      "epoch": 11.501572327044025,
      "grad_norm": 0.5079939365386963,
      "learning_rate": 2.2816550158187717e-05,
      "loss": 0.3709,
      "num_input_tokens_seen": 14354336,
      "step": 21945
    },
    {
      "epoch": 11.5041928721174,
      "grad_norm": 0.4014778435230255,
      "learning_rate": 2.2805159795446182e-05,
      "loss": 0.554,
      "num_input_tokens_seen": 14357728,
      "step": 21950
    },
    {
      "epoch": 11.506813417190775,
      "grad_norm": 0.4992315471172333,
      "learning_rate": 2.2793769891840184e-05,
      "loss": 0.4681,
      "num_input_tokens_seen": 14360384,
      "step": 21955
    },
    {
      "epoch": 11.50943396226415,
      "grad_norm": 0.40150386095046997,
      "learning_rate": 2.278238044975236e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 14364800,
      "step": 21960
    },
    {
      "epoch": 11.512054507337526,
      "grad_norm": 0.6450382471084595,
      "learning_rate": 2.2770991471565236e-05,
      "loss": 0.3117,
      "num_input_tokens_seen": 14367872,
      "step": 21965
    },
    {
      "epoch": 11.514675052410901,
      "grad_norm": 0.297756165266037,
      "learning_rate": 2.2759602959661265e-05,
      "loss": 0.5035,
      "num_input_tokens_seen": 14371872,
      "step": 21970
    },
    {
      "epoch": 11.517295597484276,
      "grad_norm": 0.8979122638702393,
      "learning_rate": 2.2748214916422796e-05,
      "loss": 0.4606,
      "num_input_tokens_seen": 14374752,
      "step": 21975
    },
    {
      "epoch": 11.519916142557651,
      "grad_norm": 1.2049199342727661,
      "learning_rate": 2.2736827344232082e-05,
      "loss": 0.4595,
      "num_input_tokens_seen": 14377792,
      "step": 21980
    },
    {
      "epoch": 11.522536687631026,
      "grad_norm": 0.28727754950523376,
      "learning_rate": 2.272544024547126e-05,
      "loss": 0.3311,
      "num_input_tokens_seen": 14382496,
      "step": 21985
    },
    {
      "epoch": 11.525157232704402,
      "grad_norm": 0.4343065321445465,
      "learning_rate": 2.27140536225224e-05,
      "loss": 0.4205,
      "num_input_tokens_seen": 14389184,
      "step": 21990
    },
    {
      "epoch": 11.527777777777779,
      "grad_norm": 0.5084301829338074,
      "learning_rate": 2.270266747776744e-05,
      "loss": 0.3949,
      "num_input_tokens_seen": 14391904,
      "step": 21995
    },
    {
      "epoch": 11.530398322851154,
      "grad_norm": 0.33370763063430786,
      "learning_rate": 2.2691281813588226e-05,
      "loss": 0.3509,
      "num_input_tokens_seen": 14394880,
      "step": 22000
    },
    {
      "epoch": 11.533018867924529,
      "grad_norm": 0.3785298764705658,
      "learning_rate": 2.267989663236652e-05,
      "loss": 0.3677,
      "num_input_tokens_seen": 14399168,
      "step": 22005
    },
    {
      "epoch": 11.535639412997904,
      "grad_norm": 0.6331139206886292,
      "learning_rate": 2.2668511936483975e-05,
      "loss": 0.6032,
      "num_input_tokens_seen": 14402848,
      "step": 22010
    },
    {
      "epoch": 11.53825995807128,
      "grad_norm": 0.3906285762786865,
      "learning_rate": 2.2657127728322124e-05,
      "loss": 0.4036,
      "num_input_tokens_seen": 14406272,
      "step": 22015
    },
    {
      "epoch": 11.540880503144654,
      "grad_norm": 0.6462644934654236,
      "learning_rate": 2.264574401026244e-05,
      "loss": 0.3993,
      "num_input_tokens_seen": 14409120,
      "step": 22020
    },
    {
      "epoch": 11.54350104821803,
      "grad_norm": 0.5035092234611511,
      "learning_rate": 2.2634360784686242e-05,
      "loss": 0.3867,
      "num_input_tokens_seen": 14412224,
      "step": 22025
    },
    {
      "epoch": 11.546121593291405,
      "grad_norm": 0.503405749797821,
      "learning_rate": 2.2622978053974776e-05,
      "loss": 0.5181,
      "num_input_tokens_seen": 14415616,
      "step": 22030
    },
    {
      "epoch": 11.54874213836478,
      "grad_norm": 0.7402815818786621,
      "learning_rate": 2.2611595820509186e-05,
      "loss": 0.693,
      "num_input_tokens_seen": 14419392,
      "step": 22035
    },
    {
      "epoch": 11.551362683438155,
      "grad_norm": 0.526427149772644,
      "learning_rate": 2.2600214086670494e-05,
      "loss": 0.3933,
      "num_input_tokens_seen": 14423040,
      "step": 22040
    },
    {
      "epoch": 11.55398322851153,
      "grad_norm": 0.44655853509902954,
      "learning_rate": 2.2588832854839644e-05,
      "loss": 0.5743,
      "num_input_tokens_seen": 14426656,
      "step": 22045
    },
    {
      "epoch": 11.556603773584905,
      "grad_norm": 1.0491093397140503,
      "learning_rate": 2.2577452127397463e-05,
      "loss": 0.5922,
      "num_input_tokens_seen": 14429792,
      "step": 22050
    },
    {
      "epoch": 11.55922431865828,
      "grad_norm": 0.7610637545585632,
      "learning_rate": 2.256607190672464e-05,
      "loss": 0.4178,
      "num_input_tokens_seen": 14432992,
      "step": 22055
    },
    {
      "epoch": 11.561844863731656,
      "grad_norm": 0.4463956356048584,
      "learning_rate": 2.2554692195201817e-05,
      "loss": 0.4733,
      "num_input_tokens_seen": 14436448,
      "step": 22060
    },
    {
      "epoch": 11.564465408805031,
      "grad_norm": 0.4901111423969269,
      "learning_rate": 2.254331299520948e-05,
      "loss": 0.4772,
      "num_input_tokens_seen": 14438880,
      "step": 22065
    },
    {
      "epoch": 11.567085953878406,
      "grad_norm": 0.7321037650108337,
      "learning_rate": 2.253193430912805e-05,
      "loss": 0.4495,
      "num_input_tokens_seen": 14441248,
      "step": 22070
    },
    {
      "epoch": 11.569706498951781,
      "grad_norm": 0.49318790435791016,
      "learning_rate": 2.2520556139337803e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 14444576,
      "step": 22075
    },
    {
      "epoch": 11.572327044025156,
      "grad_norm": 0.5229262113571167,
      "learning_rate": 2.250917848821892e-05,
      "loss": 0.4228,
      "num_input_tokens_seen": 14447968,
      "step": 22080
    },
    {
      "epoch": 11.574947589098532,
      "grad_norm": 0.2893379032611847,
      "learning_rate": 2.2497801358151472e-05,
      "loss": 0.6168,
      "num_input_tokens_seen": 14450816,
      "step": 22085
    },
    {
      "epoch": 11.577568134171909,
      "grad_norm": 0.534185528755188,
      "learning_rate": 2.248642475151545e-05,
      "loss": 0.3359,
      "num_input_tokens_seen": 14453728,
      "step": 22090
    },
    {
      "epoch": 11.580188679245284,
      "grad_norm": 0.4363027811050415,
      "learning_rate": 2.2475048670690678e-05,
      "loss": 0.5871,
      "num_input_tokens_seen": 14456512,
      "step": 22095
    },
    {
      "epoch": 11.582809224318659,
      "grad_norm": 0.48774245381355286,
      "learning_rate": 2.2463673118056927e-05,
      "loss": 0.5655,
      "num_input_tokens_seen": 14460160,
      "step": 22100
    },
    {
      "epoch": 11.585429769392034,
      "grad_norm": 0.4140772521495819,
      "learning_rate": 2.245229809599383e-05,
      "loss": 0.6057,
      "num_input_tokens_seen": 14462848,
      "step": 22105
    },
    {
      "epoch": 11.58805031446541,
      "grad_norm": 0.4590171277523041,
      "learning_rate": 2.2440923606880896e-05,
      "loss": 0.4079,
      "num_input_tokens_seen": 14466592,
      "step": 22110
    },
    {
      "epoch": 11.590670859538784,
      "grad_norm": 0.4672965705394745,
      "learning_rate": 2.242954965309755e-05,
      "loss": 0.3733,
      "num_input_tokens_seen": 14468960,
      "step": 22115
    },
    {
      "epoch": 11.59329140461216,
      "grad_norm": 0.42560020089149475,
      "learning_rate": 2.2418176237023088e-05,
      "loss": 0.3965,
      "num_input_tokens_seen": 14472864,
      "step": 22120
    },
    {
      "epoch": 11.595911949685535,
      "grad_norm": 0.3493413031101227,
      "learning_rate": 2.2406803361036706e-05,
      "loss": 0.4642,
      "num_input_tokens_seen": 14476768,
      "step": 22125
    },
    {
      "epoch": 11.59853249475891,
      "grad_norm": 0.3647328019142151,
      "learning_rate": 2.2395431027517475e-05,
      "loss": 0.3923,
      "num_input_tokens_seen": 14479872,
      "step": 22130
    },
    {
      "epoch": 11.601153039832285,
      "grad_norm": 0.4312705397605896,
      "learning_rate": 2.2384059238844364e-05,
      "loss": 0.4628,
      "num_input_tokens_seen": 14482400,
      "step": 22135
    },
    {
      "epoch": 11.60377358490566,
      "grad_norm": 0.48240137100219727,
      "learning_rate": 2.2372687997396215e-05,
      "loss": 0.4771,
      "num_input_tokens_seen": 14485696,
      "step": 22140
    },
    {
      "epoch": 11.606394129979035,
      "grad_norm": 0.4953789710998535,
      "learning_rate": 2.2361317305551756e-05,
      "loss": 0.5724,
      "num_input_tokens_seen": 14488800,
      "step": 22145
    },
    {
      "epoch": 11.60901467505241,
      "grad_norm": 0.5318878889083862,
      "learning_rate": 2.2349947165689615e-05,
      "loss": 0.4023,
      "num_input_tokens_seen": 14491520,
      "step": 22150
    },
    {
      "epoch": 11.611635220125786,
      "grad_norm": 0.4273078441619873,
      "learning_rate": 2.2338577580188296e-05,
      "loss": 0.3679,
      "num_input_tokens_seen": 14494176,
      "step": 22155
    },
    {
      "epoch": 11.614255765199161,
      "grad_norm": 0.6334779858589172,
      "learning_rate": 2.2327208551426188e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 14496640,
      "step": 22160
    },
    {
      "epoch": 11.616876310272536,
      "grad_norm": 0.7594039440155029,
      "learning_rate": 2.2315840081781546e-05,
      "loss": 0.4302,
      "num_input_tokens_seen": 14499232,
      "step": 22165
    },
    {
      "epoch": 11.619496855345911,
      "grad_norm": 0.6527590155601501,
      "learning_rate": 2.230447217363255e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 14501600,
      "step": 22170
    },
    {
      "epoch": 11.622117400419286,
      "grad_norm": 0.3307563066482544,
      "learning_rate": 2.2293104829357212e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 14504544,
      "step": 22175
    },
    {
      "epoch": 11.624737945492662,
      "grad_norm": 1.0196391344070435,
      "learning_rate": 2.2281738051333467e-05,
      "loss": 0.3646,
      "num_input_tokens_seen": 14507904,
      "step": 22180
    },
    {
      "epoch": 11.627358490566039,
      "grad_norm": 0.2707454562187195,
      "learning_rate": 2.2270371841939105e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 14511264,
      "step": 22185
    },
    {
      "epoch": 11.629979035639414,
      "grad_norm": 0.33909955620765686,
      "learning_rate": 2.2259006203551824e-05,
      "loss": 0.4034,
      "num_input_tokens_seen": 14514752,
      "step": 22190
    },
    {
      "epoch": 11.632599580712789,
      "grad_norm": 0.3790900707244873,
      "learning_rate": 2.2247641138549168e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 14517312,
      "step": 22195
    },
    {
      "epoch": 11.635220125786164,
      "grad_norm": 0.4695182144641876,
      "learning_rate": 2.2236276649308583e-05,
      "loss": 0.442,
      "num_input_tokens_seen": 14520800,
      "step": 22200
    },
    {
      "epoch": 11.63784067085954,
      "grad_norm": 0.40549546480178833,
      "learning_rate": 2.2224912738207395e-05,
      "loss": 0.3785,
      "num_input_tokens_seen": 14524000,
      "step": 22205
    },
    {
      "epoch": 11.640461215932914,
      "grad_norm": 0.2659561038017273,
      "learning_rate": 2.2213549407622796e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 14526944,
      "step": 22210
    },
    {
      "epoch": 11.64308176100629,
      "grad_norm": 0.4158248007297516,
      "learning_rate": 2.220218665993188e-05,
      "loss": 0.338,
      "num_input_tokens_seen": 14530368,
      "step": 22215
    },
    {
      "epoch": 11.645702306079665,
      "grad_norm": 0.4348396062850952,
      "learning_rate": 2.21908244975116e-05,
      "loss": 0.3787,
      "num_input_tokens_seen": 14533824,
      "step": 22220
    },
    {
      "epoch": 11.64832285115304,
      "grad_norm": 0.769169270992279,
      "learning_rate": 2.2179462922738776e-05,
      "loss": 0.3779,
      "num_input_tokens_seen": 14539232,
      "step": 22225
    },
    {
      "epoch": 11.650943396226415,
      "grad_norm": 0.3452356159687042,
      "learning_rate": 2.2168101937990135e-05,
      "loss": 0.3885,
      "num_input_tokens_seen": 14542784,
      "step": 22230
    },
    {
      "epoch": 11.65356394129979,
      "grad_norm": 0.5404634475708008,
      "learning_rate": 2.2156741545642253e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 14545696,
      "step": 22235
    },
    {
      "epoch": 11.656184486373165,
      "grad_norm": 0.43357253074645996,
      "learning_rate": 2.2145381748071606e-05,
      "loss": 0.2801,
      "num_input_tokens_seen": 14549344,
      "step": 22240
    },
    {
      "epoch": 11.65880503144654,
      "grad_norm": 0.3845861852169037,
      "learning_rate": 2.2134022547654538e-05,
      "loss": 0.4361,
      "num_input_tokens_seen": 14552096,
      "step": 22245
    },
    {
      "epoch": 11.661425576519916,
      "grad_norm": 0.661048173904419,
      "learning_rate": 2.212266394676725e-05,
      "loss": 0.3491,
      "num_input_tokens_seen": 14556384,
      "step": 22250
    },
    {
      "epoch": 11.664046121593291,
      "grad_norm": 0.5720515251159668,
      "learning_rate": 2.2111305947785836e-05,
      "loss": 0.4136,
      "num_input_tokens_seen": 14559680,
      "step": 22255
    },
    {
      "epoch": 11.666666666666666,
      "grad_norm": 0.4215748906135559,
      "learning_rate": 2.2099948553086263e-05,
      "loss": 0.3737,
      "num_input_tokens_seen": 14562592,
      "step": 22260
    },
    {
      "epoch": 11.669287211740041,
      "grad_norm": 0.5479528307914734,
      "learning_rate": 2.2088591765044365e-05,
      "loss": 0.5232,
      "num_input_tokens_seen": 14564960,
      "step": 22265
    },
    {
      "epoch": 11.671907756813416,
      "grad_norm": 0.48211976885795593,
      "learning_rate": 2.207723558603586e-05,
      "loss": 0.4898,
      "num_input_tokens_seen": 14568384,
      "step": 22270
    },
    {
      "epoch": 11.674528301886792,
      "grad_norm": 0.45490506291389465,
      "learning_rate": 2.206588001843633e-05,
      "loss": 0.4106,
      "num_input_tokens_seen": 14571680,
      "step": 22275
    },
    {
      "epoch": 11.677148846960169,
      "grad_norm": 0.6486725807189941,
      "learning_rate": 2.2054525064621216e-05,
      "loss": 0.4645,
      "num_input_tokens_seen": 14575136,
      "step": 22280
    },
    {
      "epoch": 11.679769392033544,
      "grad_norm": 0.500831663608551,
      "learning_rate": 2.2043170726965858e-05,
      "loss": 0.3953,
      "num_input_tokens_seen": 14577888,
      "step": 22285
    },
    {
      "epoch": 11.682389937106919,
      "grad_norm": 0.34825563430786133,
      "learning_rate": 2.2031817007845447e-05,
      "loss": 0.4067,
      "num_input_tokens_seen": 14581248,
      "step": 22290
    },
    {
      "epoch": 11.685010482180294,
      "grad_norm": 0.6307078003883362,
      "learning_rate": 2.202046390963506e-05,
      "loss": 0.5075,
      "num_input_tokens_seen": 14584928,
      "step": 22295
    },
    {
      "epoch": 11.68763102725367,
      "grad_norm": 0.9108502268791199,
      "learning_rate": 2.2009111434709627e-05,
      "loss": 0.6013,
      "num_input_tokens_seen": 14587552,
      "step": 22300
    },
    {
      "epoch": 11.690251572327044,
      "grad_norm": 0.6383960843086243,
      "learning_rate": 2.1997759585443968e-05,
      "loss": 0.4391,
      "num_input_tokens_seen": 14589760,
      "step": 22305
    },
    {
      "epoch": 11.69287211740042,
      "grad_norm": 0.6917334198951721,
      "learning_rate": 2.198640836421275e-05,
      "loss": 0.507,
      "num_input_tokens_seen": 14592448,
      "step": 22310
    },
    {
      "epoch": 11.695492662473795,
      "grad_norm": 0.7545520067214966,
      "learning_rate": 2.1975057773390515e-05,
      "loss": 0.4523,
      "num_input_tokens_seen": 14595808,
      "step": 22315
    },
    {
      "epoch": 11.69811320754717,
      "grad_norm": 0.41791149973869324,
      "learning_rate": 2.1963707815351683e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 14598624,
      "step": 22320
    },
    {
      "epoch": 11.700733752620545,
      "grad_norm": 1.1504063606262207,
      "learning_rate": 2.1952358492470544e-05,
      "loss": 0.406,
      "num_input_tokens_seen": 14601536,
      "step": 22325
    },
    {
      "epoch": 11.70335429769392,
      "grad_norm": 0.4138602018356323,
      "learning_rate": 2.194100980712124e-05,
      "loss": 0.4343,
      "num_input_tokens_seen": 14604736,
      "step": 22330
    },
    {
      "epoch": 11.705974842767295,
      "grad_norm": 0.540610671043396,
      "learning_rate": 2.1929661761677774e-05,
      "loss": 0.3859,
      "num_input_tokens_seen": 14607840,
      "step": 22335
    },
    {
      "epoch": 11.70859538784067,
      "grad_norm": 0.5313720703125,
      "learning_rate": 2.191831435851405e-05,
      "loss": 0.5085,
      "num_input_tokens_seen": 14610816,
      "step": 22340
    },
    {
      "epoch": 11.711215932914046,
      "grad_norm": 0.5753461718559265,
      "learning_rate": 2.1906967600003793e-05,
      "loss": 0.5113,
      "num_input_tokens_seen": 14614464,
      "step": 22345
    },
    {
      "epoch": 11.713836477987421,
      "grad_norm": 0.4351661205291748,
      "learning_rate": 2.1895621488520636e-05,
      "loss": 0.3888,
      "num_input_tokens_seen": 14616704,
      "step": 22350
    },
    {
      "epoch": 11.716457023060796,
      "grad_norm": 0.33827486634254456,
      "learning_rate": 2.1884276026438044e-05,
      "loss": 0.4552,
      "num_input_tokens_seen": 14620832,
      "step": 22355
    },
    {
      "epoch": 11.719077568134171,
      "grad_norm": 0.4660438895225525,
      "learning_rate": 2.187293121612937e-05,
      "loss": 0.4923,
      "num_input_tokens_seen": 14623328,
      "step": 22360
    },
    {
      "epoch": 11.721698113207546,
      "grad_norm": 0.3043864071369171,
      "learning_rate": 2.1861587059967804e-05,
      "loss": 0.4559,
      "num_input_tokens_seen": 14627520,
      "step": 22365
    },
    {
      "epoch": 11.724318658280922,
      "grad_norm": 0.798977792263031,
      "learning_rate": 2.1850243560326414e-05,
      "loss": 0.3876,
      "num_input_tokens_seen": 14631712,
      "step": 22370
    },
    {
      "epoch": 11.726939203354299,
      "grad_norm": 0.3928571045398712,
      "learning_rate": 2.1838900719578143e-05,
      "loss": 0.456,
      "num_input_tokens_seen": 14634912,
      "step": 22375
    },
    {
      "epoch": 11.729559748427674,
      "grad_norm": 0.504433274269104,
      "learning_rate": 2.1827558540095772e-05,
      "loss": 0.4872,
      "num_input_tokens_seen": 14637440,
      "step": 22380
    },
    {
      "epoch": 11.732180293501049,
      "grad_norm": 0.19810019433498383,
      "learning_rate": 2.1816217024251966e-05,
      "loss": 0.4578,
      "num_input_tokens_seen": 14642336,
      "step": 22385
    },
    {
      "epoch": 11.734800838574424,
      "grad_norm": 0.6453500986099243,
      "learning_rate": 2.1804876174419242e-05,
      "loss": 0.4803,
      "num_input_tokens_seen": 14646016,
      "step": 22390
    },
    {
      "epoch": 11.7374213836478,
      "grad_norm": 0.3919985592365265,
      "learning_rate": 2.179353599296996e-05,
      "loss": 0.3816,
      "num_input_tokens_seen": 14648384,
      "step": 22395
    },
    {
      "epoch": 11.740041928721174,
      "grad_norm": 0.4276735484600067,
      "learning_rate": 2.1782196482276368e-05,
      "loss": 0.5112,
      "num_input_tokens_seen": 14651552,
      "step": 22400
    },
    {
      "epoch": 11.74266247379455,
      "grad_norm": 0.36120620369911194,
      "learning_rate": 2.1770857644710567e-05,
      "loss": 0.5541,
      "num_input_tokens_seen": 14654304,
      "step": 22405
    },
    {
      "epoch": 11.745283018867925,
      "grad_norm": 0.5650772452354431,
      "learning_rate": 2.1759519482644495e-05,
      "loss": 0.4328,
      "num_input_tokens_seen": 14657472,
      "step": 22410
    },
    {
      "epoch": 11.7479035639413,
      "grad_norm": 0.4333966374397278,
      "learning_rate": 2.174818199844999e-05,
      "loss": 0.5947,
      "num_input_tokens_seen": 14660992,
      "step": 22415
    },
    {
      "epoch": 11.750524109014675,
      "grad_norm": 0.32627052068710327,
      "learning_rate": 2.173684519449872e-05,
      "loss": 0.4526,
      "num_input_tokens_seen": 14665344,
      "step": 22420
    },
    {
      "epoch": 11.75314465408805,
      "grad_norm": 0.5436223745346069,
      "learning_rate": 2.1725509073162188e-05,
      "loss": 0.5261,
      "num_input_tokens_seen": 14668384,
      "step": 22425
    },
    {
      "epoch": 11.755765199161425,
      "grad_norm": 0.39873912930488586,
      "learning_rate": 2.171417363681181e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 14671488,
      "step": 22430
    },
    {
      "epoch": 11.7583857442348,
      "grad_norm": 0.873200535774231,
      "learning_rate": 2.1702838887818816e-05,
      "loss": 0.4987,
      "num_input_tokens_seen": 14674592,
      "step": 22435
    },
    {
      "epoch": 11.761006289308176,
      "grad_norm": 0.38557693362236023,
      "learning_rate": 2.1691504828554315e-05,
      "loss": 0.5529,
      "num_input_tokens_seen": 14678176,
      "step": 22440
    },
    {
      "epoch": 11.76362683438155,
      "grad_norm": 0.42651280760765076,
      "learning_rate": 2.1680171461389265e-05,
      "loss": 0.385,
      "num_input_tokens_seen": 14681184,
      "step": 22445
    },
    {
      "epoch": 11.766247379454926,
      "grad_norm": 0.27146902680397034,
      "learning_rate": 2.166883878869446e-05,
      "loss": 0.356,
      "num_input_tokens_seen": 14684192,
      "step": 22450
    },
    {
      "epoch": 11.768867924528301,
      "grad_norm": 0.4567176103591919,
      "learning_rate": 2.1657506812840582e-05,
      "loss": 0.4086,
      "num_input_tokens_seen": 14687040,
      "step": 22455
    },
    {
      "epoch": 11.771488469601676,
      "grad_norm": 0.2756119966506958,
      "learning_rate": 2.164617553619814e-05,
      "loss": 0.5681,
      "num_input_tokens_seen": 14690592,
      "step": 22460
    },
    {
      "epoch": 11.774109014675052,
      "grad_norm": 0.41460272669792175,
      "learning_rate": 2.1634844961137517e-05,
      "loss": 0.4091,
      "num_input_tokens_seen": 14693792,
      "step": 22465
    },
    {
      "epoch": 11.776729559748428,
      "grad_norm": 0.6452013254165649,
      "learning_rate": 2.162351509002893e-05,
      "loss": 0.622,
      "num_input_tokens_seen": 14696544,
      "step": 22470
    },
    {
      "epoch": 11.779350104821804,
      "grad_norm": 0.39064428210258484,
      "learning_rate": 2.161218592524248e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 14700416,
      "step": 22475
    },
    {
      "epoch": 11.781970649895179,
      "grad_norm": 0.5480527877807617,
      "learning_rate": 2.160085746914807e-05,
      "loss": 0.5411,
      "num_input_tokens_seen": 14703648,
      "step": 22480
    },
    {
      "epoch": 11.784591194968554,
      "grad_norm": 0.6128780841827393,
      "learning_rate": 2.1589529724115503e-05,
      "loss": 0.4063,
      "num_input_tokens_seen": 14706304,
      "step": 22485
    },
    {
      "epoch": 11.78721174004193,
      "grad_norm": 0.4292333126068115,
      "learning_rate": 2.1578202692514407e-05,
      "loss": 0.4726,
      "num_input_tokens_seen": 14708672,
      "step": 22490
    },
    {
      "epoch": 11.789832285115304,
      "grad_norm": 0.642744243144989,
      "learning_rate": 2.1566876376714274e-05,
      "loss": 0.57,
      "num_input_tokens_seen": 14711904,
      "step": 22495
    },
    {
      "epoch": 11.79245283018868,
      "grad_norm": 1.0713353157043457,
      "learning_rate": 2.155555077908443e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 14714080,
      "step": 22500
    },
    {
      "epoch": 11.795073375262055,
      "grad_norm": 0.6458566188812256,
      "learning_rate": 2.1544225901994077e-05,
      "loss": 0.358,
      "num_input_tokens_seen": 14717152,
      "step": 22505
    },
    {
      "epoch": 11.79769392033543,
      "grad_norm": 0.42851176857948303,
      "learning_rate": 2.1532901747812234e-05,
      "loss": 0.4016,
      "num_input_tokens_seen": 14720064,
      "step": 22510
    },
    {
      "epoch": 11.800314465408805,
      "grad_norm": 0.48194795846939087,
      "learning_rate": 2.1521578318907787e-05,
      "loss": 0.5051,
      "num_input_tokens_seen": 14722848,
      "step": 22515
    },
    {
      "epoch": 11.80293501048218,
      "grad_norm": 0.48053234815597534,
      "learning_rate": 2.1510255617649476e-05,
      "loss": 0.3707,
      "num_input_tokens_seen": 14725696,
      "step": 22520
    },
    {
      "epoch": 11.805555555555555,
      "grad_norm": 0.6309667825698853,
      "learning_rate": 2.149893364640587e-05,
      "loss": 0.4355,
      "num_input_tokens_seen": 14728800,
      "step": 22525
    },
    {
      "epoch": 11.80817610062893,
      "grad_norm": 0.30860376358032227,
      "learning_rate": 2.1487612407545417e-05,
      "loss": 0.3441,
      "num_input_tokens_seen": 14732608,
      "step": 22530
    },
    {
      "epoch": 11.810796645702306,
      "grad_norm": 0.39198148250579834,
      "learning_rate": 2.147629190343637e-05,
      "loss": 0.4339,
      "num_input_tokens_seen": 14736640,
      "step": 22535
    },
    {
      "epoch": 11.81341719077568,
      "grad_norm": 0.4607129395008087,
      "learning_rate": 2.1464972136446852e-05,
      "loss": 0.3359,
      "num_input_tokens_seen": 14739296,
      "step": 22540
    },
    {
      "epoch": 11.816037735849056,
      "grad_norm": 0.47945842146873474,
      "learning_rate": 2.1453653108944838e-05,
      "loss": 0.3365,
      "num_input_tokens_seen": 14742048,
      "step": 22545
    },
    {
      "epoch": 11.818658280922431,
      "grad_norm": 0.7632633447647095,
      "learning_rate": 2.1442334823298128e-05,
      "loss": 0.3684,
      "num_input_tokens_seen": 14745056,
      "step": 22550
    },
    {
      "epoch": 11.821278825995806,
      "grad_norm": 0.4764186441898346,
      "learning_rate": 2.1431017281874386e-05,
      "loss": 0.6168,
      "num_input_tokens_seen": 14748064,
      "step": 22555
    },
    {
      "epoch": 11.823899371069182,
      "grad_norm": 0.5429674386978149,
      "learning_rate": 2.1419700487041123e-05,
      "loss": 0.5072,
      "num_input_tokens_seen": 14751264,
      "step": 22560
    },
    {
      "epoch": 11.826519916142558,
      "grad_norm": 0.3528273403644562,
      "learning_rate": 2.140838444116566e-05,
      "loss": 0.5368,
      "num_input_tokens_seen": 14755168,
      "step": 22565
    },
    {
      "epoch": 11.829140461215934,
      "grad_norm": 0.982427179813385,
      "learning_rate": 2.13970691466152e-05,
      "loss": 0.5945,
      "num_input_tokens_seen": 14757408,
      "step": 22570
    },
    {
      "epoch": 11.831761006289309,
      "grad_norm": 0.3890267312526703,
      "learning_rate": 2.1385754605756773e-05,
      "loss": 0.4823,
      "num_input_tokens_seen": 14760288,
      "step": 22575
    },
    {
      "epoch": 11.834381551362684,
      "grad_norm": 0.4991300702095032,
      "learning_rate": 2.137444082095724e-05,
      "loss": 0.4523,
      "num_input_tokens_seen": 14763616,
      "step": 22580
    },
    {
      "epoch": 11.83700209643606,
      "grad_norm": 0.713006854057312,
      "learning_rate": 2.1363127794583335e-05,
      "loss": 0.3857,
      "num_input_tokens_seen": 14768544,
      "step": 22585
    },
    {
      "epoch": 11.839622641509434,
      "grad_norm": 0.4856227934360504,
      "learning_rate": 2.1351815529001605e-05,
      "loss": 0.5315,
      "num_input_tokens_seen": 14771456,
      "step": 22590
    },
    {
      "epoch": 11.84224318658281,
      "grad_norm": 0.3476814031600952,
      "learning_rate": 2.1340504026578438e-05,
      "loss": 0.5487,
      "num_input_tokens_seen": 14774880,
      "step": 22595
    },
    {
      "epoch": 11.844863731656185,
      "grad_norm": 0.41603100299835205,
      "learning_rate": 2.1329193289680084e-05,
      "loss": 0.418,
      "num_input_tokens_seen": 14778432,
      "step": 22600
    },
    {
      "epoch": 11.84748427672956,
      "grad_norm": 0.6406997442245483,
      "learning_rate": 2.1317883320672608e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 14781984,
      "step": 22605
    },
    {
      "epoch": 11.850104821802935,
      "grad_norm": 0.8305124044418335,
      "learning_rate": 2.1306574121921942e-05,
      "loss": 0.5399,
      "num_input_tokens_seen": 14785280,
      "step": 22610
    },
    {
      "epoch": 11.85272536687631,
      "grad_norm": 0.5543068647384644,
      "learning_rate": 2.1295265695793836e-05,
      "loss": 0.5166,
      "num_input_tokens_seen": 14789568,
      "step": 22615
    },
    {
      "epoch": 11.855345911949685,
      "grad_norm": 0.592064380645752,
      "learning_rate": 2.128395804465387e-05,
      "loss": 0.3464,
      "num_input_tokens_seen": 14792608,
      "step": 22620
    },
    {
      "epoch": 11.85796645702306,
      "grad_norm": 0.11316777765750885,
      "learning_rate": 2.1272651170867487e-05,
      "loss": 0.3214,
      "num_input_tokens_seen": 14797760,
      "step": 22625
    },
    {
      "epoch": 11.860587002096436,
      "grad_norm": 0.4504036605358124,
      "learning_rate": 2.1261345076799955e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 14800768,
      "step": 22630
    },
    {
      "epoch": 11.86320754716981,
      "grad_norm": 0.23604463040828705,
      "learning_rate": 2.125003976481638e-05,
      "loss": 0.4211,
      "num_input_tokens_seen": 14804000,
      "step": 22635
    },
    {
      "epoch": 11.865828092243186,
      "grad_norm": 0.4019843339920044,
      "learning_rate": 2.123873523728171e-05,
      "loss": 0.327,
      "num_input_tokens_seen": 14807392,
      "step": 22640
    },
    {
      "epoch": 11.868448637316561,
      "grad_norm": 0.3961583375930786,
      "learning_rate": 2.122743149656072e-05,
      "loss": 0.4076,
      "num_input_tokens_seen": 14810816,
      "step": 22645
    },
    {
      "epoch": 11.871069182389936,
      "grad_norm": 0.6408613324165344,
      "learning_rate": 2.1216128545018012e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 14813056,
      "step": 22650
    },
    {
      "epoch": 11.873689727463312,
      "grad_norm": 0.23569050431251526,
      "learning_rate": 2.1204826385018052e-05,
      "loss": 0.4817,
      "num_input_tokens_seen": 14816736,
      "step": 22655
    },
    {
      "epoch": 11.876310272536688,
      "grad_norm": 0.33024275302886963,
      "learning_rate": 2.1193525018925106e-05,
      "loss": 0.5112,
      "num_input_tokens_seen": 14820160,
      "step": 22660
    },
    {
      "epoch": 11.878930817610064,
      "grad_norm": 0.43698760867118835,
      "learning_rate": 2.1182224449103304e-05,
      "loss": 0.4558,
      "num_input_tokens_seen": 14822944,
      "step": 22665
    },
    {
      "epoch": 11.881551362683439,
      "grad_norm": 0.6238161325454712,
      "learning_rate": 2.117092467791659e-05,
      "loss": 0.4207,
      "num_input_tokens_seen": 14826464,
      "step": 22670
    },
    {
      "epoch": 11.884171907756814,
      "grad_norm": 1.0937429666519165,
      "learning_rate": 2.115962570772876e-05,
      "loss": 0.4288,
      "num_input_tokens_seen": 14830080,
      "step": 22675
    },
    {
      "epoch": 11.88679245283019,
      "grad_norm": 0.39909589290618896,
      "learning_rate": 2.1148327540903408e-05,
      "loss": 0.4819,
      "num_input_tokens_seen": 14833280,
      "step": 22680
    },
    {
      "epoch": 11.889412997903564,
      "grad_norm": 0.4484791159629822,
      "learning_rate": 2.113703017980399e-05,
      "loss": 0.5633,
      "num_input_tokens_seen": 14836768,
      "step": 22685
    },
    {
      "epoch": 11.89203354297694,
      "grad_norm": 0.7098581790924072,
      "learning_rate": 2.1125733626793793e-05,
      "loss": 0.5314,
      "num_input_tokens_seen": 14839456,
      "step": 22690
    },
    {
      "epoch": 11.894654088050315,
      "grad_norm": 0.6414666175842285,
      "learning_rate": 2.1114437884235915e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 14842272,
      "step": 22695
    },
    {
      "epoch": 11.89727463312369,
      "grad_norm": 0.39460331201553345,
      "learning_rate": 2.1103142954493313e-05,
      "loss": 0.5964,
      "num_input_tokens_seen": 14846304,
      "step": 22700
    },
    {
      "epoch": 11.899895178197065,
      "grad_norm": 0.5928078293800354,
      "learning_rate": 2.1091848839928745e-05,
      "loss": 0.4622,
      "num_input_tokens_seen": 14849184,
      "step": 22705
    },
    {
      "epoch": 11.90251572327044,
      "grad_norm": 0.6447815895080566,
      "learning_rate": 2.1080555542904805e-05,
      "loss": 0.419,
      "num_input_tokens_seen": 14852704,
      "step": 22710
    },
    {
      "epoch": 11.905136268343815,
      "grad_norm": 0.35902753472328186,
      "learning_rate": 2.106926306578393e-05,
      "loss": 0.3926,
      "num_input_tokens_seen": 14856288,
      "step": 22715
    },
    {
      "epoch": 11.90775681341719,
      "grad_norm": 0.740516722202301,
      "learning_rate": 2.1057971410928383e-05,
      "loss": 0.4262,
      "num_input_tokens_seen": 14859232,
      "step": 22720
    },
    {
      "epoch": 11.910377358490566,
      "grad_norm": 0.5112674236297607,
      "learning_rate": 2.104668058070024e-05,
      "loss": 0.4344,
      "num_input_tokens_seen": 14862624,
      "step": 22725
    },
    {
      "epoch": 11.91299790356394,
      "grad_norm": 0.6987025141716003,
      "learning_rate": 2.1035390577461427e-05,
      "loss": 0.4649,
      "num_input_tokens_seen": 14866368,
      "step": 22730
    },
    {
      "epoch": 11.915618448637316,
      "grad_norm": 0.3801637589931488,
      "learning_rate": 2.102410140357367e-05,
      "loss": 0.3527,
      "num_input_tokens_seen": 14869024,
      "step": 22735
    },
    {
      "epoch": 11.918238993710691,
      "grad_norm": 0.39983177185058594,
      "learning_rate": 2.1012813061398534e-05,
      "loss": 0.3643,
      "num_input_tokens_seen": 14871936,
      "step": 22740
    },
    {
      "epoch": 11.920859538784066,
      "grad_norm": 0.5205259323120117,
      "learning_rate": 2.1001525553297422e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 14874400,
      "step": 22745
    },
    {
      "epoch": 11.923480083857442,
      "grad_norm": 0.5398586392402649,
      "learning_rate": 2.099023888163154e-05,
      "loss": 0.3973,
      "num_input_tokens_seen": 14876864,
      "step": 22750
    },
    {
      "epoch": 11.926100628930818,
      "grad_norm": 0.4887113571166992,
      "learning_rate": 2.0978953048761947e-05,
      "loss": 0.3789,
      "num_input_tokens_seen": 14879872,
      "step": 22755
    },
    {
      "epoch": 11.928721174004194,
      "grad_norm": 0.43427062034606934,
      "learning_rate": 2.0967668057049507e-05,
      "loss": 0.4464,
      "num_input_tokens_seen": 14883424,
      "step": 22760
    },
    {
      "epoch": 11.931341719077569,
      "grad_norm": 0.5546324253082275,
      "learning_rate": 2.0956383908854895e-05,
      "loss": 0.3966,
      "num_input_tokens_seen": 14887392,
      "step": 22765
    },
    {
      "epoch": 11.933962264150944,
      "grad_norm": 0.48177874088287354,
      "learning_rate": 2.0945100606538644e-05,
      "loss": 0.5753,
      "num_input_tokens_seen": 14890688,
      "step": 22770
    },
    {
      "epoch": 11.93658280922432,
      "grad_norm": 0.3417784869670868,
      "learning_rate": 2.0933818152461077e-05,
      "loss": 0.3129,
      "num_input_tokens_seen": 14894464,
      "step": 22775
    },
    {
      "epoch": 11.939203354297694,
      "grad_norm": 0.6113790273666382,
      "learning_rate": 2.092253654898237e-05,
      "loss": 0.4337,
      "num_input_tokens_seen": 14897024,
      "step": 22780
    },
    {
      "epoch": 11.94182389937107,
      "grad_norm": 0.373384028673172,
      "learning_rate": 2.0911255798462494e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 14900480,
      "step": 22785
    },
    {
      "epoch": 11.944444444444445,
      "grad_norm": 0.7278722524642944,
      "learning_rate": 2.0899975903261266e-05,
      "loss": 0.5136,
      "num_input_tokens_seen": 14903136,
      "step": 22790
    },
    {
      "epoch": 11.94706498951782,
      "grad_norm": 0.43077653646469116,
      "learning_rate": 2.0888696865738295e-05,
      "loss": 0.3924,
      "num_input_tokens_seen": 14907040,
      "step": 22795
    },
    {
      "epoch": 11.949685534591195,
      "grad_norm": 0.501041829586029,
      "learning_rate": 2.087741868825304e-05,
      "loss": 0.599,
      "num_input_tokens_seen": 14910304,
      "step": 22800
    },
    {
      "epoch": 11.95230607966457,
      "grad_norm": 0.35210704803466797,
      "learning_rate": 2.0866141373164758e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 14913312,
      "step": 22805
    },
    {
      "epoch": 11.954926624737945,
      "grad_norm": 0.4622713625431061,
      "learning_rate": 2.0854864922832544e-05,
      "loss": 0.5503,
      "num_input_tokens_seen": 14916512,
      "step": 22810
    },
    {
      "epoch": 11.95754716981132,
      "grad_norm": 0.3868999779224396,
      "learning_rate": 2.0843589339615304e-05,
      "loss": 0.4241,
      "num_input_tokens_seen": 14920416,
      "step": 22815
    },
    {
      "epoch": 11.960167714884696,
      "grad_norm": 0.4793449342250824,
      "learning_rate": 2.0832314625871742e-05,
      "loss": 0.5062,
      "num_input_tokens_seen": 14923840,
      "step": 22820
    },
    {
      "epoch": 11.96278825995807,
      "grad_norm": 0.41069647669792175,
      "learning_rate": 2.0821040783960423e-05,
      "loss": 0.4729,
      "num_input_tokens_seen": 14927520,
      "step": 22825
    },
    {
      "epoch": 11.965408805031446,
      "grad_norm": 0.6929935812950134,
      "learning_rate": 2.0809767816239686e-05,
      "loss": 0.3778,
      "num_input_tokens_seen": 14930240,
      "step": 22830
    },
    {
      "epoch": 11.968029350104821,
      "grad_norm": 0.6548711657524109,
      "learning_rate": 2.0798495725067725e-05,
      "loss": 0.4436,
      "num_input_tokens_seen": 14932928,
      "step": 22835
    },
    {
      "epoch": 11.970649895178196,
      "grad_norm": 0.8043416738510132,
      "learning_rate": 2.078722451280252e-05,
      "loss": 0.4547,
      "num_input_tokens_seen": 14935616,
      "step": 22840
    },
    {
      "epoch": 11.973270440251572,
      "grad_norm": 0.5283016562461853,
      "learning_rate": 2.0775954181801893e-05,
      "loss": 0.398,
      "num_input_tokens_seen": 14938080,
      "step": 22845
    },
    {
      "epoch": 11.975890985324948,
      "grad_norm": 1.2367453575134277,
      "learning_rate": 2.0764684734423455e-05,
      "loss": 0.5763,
      "num_input_tokens_seen": 14940320,
      "step": 22850
    },
    {
      "epoch": 11.978511530398324,
      "grad_norm": 0.6803041696548462,
      "learning_rate": 2.075341617302465e-05,
      "loss": 0.4651,
      "num_input_tokens_seen": 14944000,
      "step": 22855
    },
    {
      "epoch": 11.981132075471699,
      "grad_norm": 0.33145639300346375,
      "learning_rate": 2.074214849996274e-05,
      "loss": 0.3638,
      "num_input_tokens_seen": 14947168,
      "step": 22860
    },
    {
      "epoch": 11.983752620545074,
      "grad_norm": 0.3740949034690857,
      "learning_rate": 2.0730881717594776e-05,
      "loss": 0.4483,
      "num_input_tokens_seen": 14950592,
      "step": 22865
    },
    {
      "epoch": 11.98637316561845,
      "grad_norm": 0.5899195075035095,
      "learning_rate": 2.0719615828277657e-05,
      "loss": 0.4252,
      "num_input_tokens_seen": 14954176,
      "step": 22870
    },
    {
      "epoch": 11.988993710691824,
      "grad_norm": 0.7696555852890015,
      "learning_rate": 2.0708350834368083e-05,
      "loss": 0.472,
      "num_input_tokens_seen": 14957248,
      "step": 22875
    },
    {
      "epoch": 11.9916142557652,
      "grad_norm": 0.4570530354976654,
      "learning_rate": 2.069708673822255e-05,
      "loss": 0.4948,
      "num_input_tokens_seen": 14960640,
      "step": 22880
    },
    {
      "epoch": 11.994234800838575,
      "grad_norm": 0.6195477843284607,
      "learning_rate": 2.068582354219737e-05,
      "loss": 0.4105,
      "num_input_tokens_seen": 14963392,
      "step": 22885
    },
    {
      "epoch": 11.99685534591195,
      "grad_norm": 0.3406940698623657,
      "learning_rate": 2.0674561248648697e-05,
      "loss": 0.4411,
      "num_input_tokens_seen": 14966528,
      "step": 22890
    },
    {
      "epoch": 11.999475890985325,
      "grad_norm": 0.6169590950012207,
      "learning_rate": 2.0663299859932456e-05,
      "loss": 0.4424,
      "num_input_tokens_seen": 14970464,
      "step": 22895
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.4754877984523773,
      "eval_runtime": 9.7271,
      "eval_samples_per_second": 87.179,
      "eval_steps_per_second": 21.795,
      "num_input_tokens_seen": 14970528,
      "step": 22896
    },
    {
      "epoch": 12.0020964360587,
      "grad_norm": 0.5609192252159119,
      "learning_rate": 2.065203937840442e-05,
      "loss": 0.401,
      "num_input_tokens_seen": 14973600,
      "step": 22900
    },
    {
      "epoch": 12.004716981132075,
      "grad_norm": 0.32814842462539673,
      "learning_rate": 2.0640779806420136e-05,
      "loss": 0.5247,
      "num_input_tokens_seen": 14977536,
      "step": 22905
    },
    {
      "epoch": 12.00733752620545,
      "grad_norm": 0.2717117667198181,
      "learning_rate": 2.0629521146334978e-05,
      "loss": 0.3522,
      "num_input_tokens_seen": 14980640,
      "step": 22910
    },
    {
      "epoch": 12.009958071278826,
      "grad_norm": 0.5024399161338806,
      "learning_rate": 2.061826340050414e-05,
      "loss": 0.4079,
      "num_input_tokens_seen": 14983456,
      "step": 22915
    },
    {
      "epoch": 12.0125786163522,
      "grad_norm": 0.4729185998439789,
      "learning_rate": 2.0607006571282605e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 14987328,
      "step": 22920
    },
    {
      "epoch": 12.015199161425576,
      "grad_norm": 0.7786957025527954,
      "learning_rate": 2.0595750661025183e-05,
      "loss": 0.4146,
      "num_input_tokens_seen": 14990432,
      "step": 22925
    },
    {
      "epoch": 12.017819706498951,
      "grad_norm": 0.4782746434211731,
      "learning_rate": 2.0584495672086483e-05,
      "loss": 0.4709,
      "num_input_tokens_seen": 14993824,
      "step": 22930
    },
    {
      "epoch": 12.020440251572326,
      "grad_norm": 0.4072149991989136,
      "learning_rate": 2.05732416068209e-05,
      "loss": 0.4005,
      "num_input_tokens_seen": 14996608,
      "step": 22935
    },
    {
      "epoch": 12.023060796645701,
      "grad_norm": 0.3033732771873474,
      "learning_rate": 2.0561988467582676e-05,
      "loss": 0.4016,
      "num_input_tokens_seen": 15000128,
      "step": 22940
    },
    {
      "epoch": 12.025681341719078,
      "grad_norm": 0.5201624035835266,
      "learning_rate": 2.0550736256725828e-05,
      "loss": 0.4583,
      "num_input_tokens_seen": 15003104,
      "step": 22945
    },
    {
      "epoch": 12.028301886792454,
      "grad_norm": 0.6019538044929504,
      "learning_rate": 2.0539484976604196e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 15005632,
      "step": 22950
    },
    {
      "epoch": 12.030922431865829,
      "grad_norm": 0.572091817855835,
      "learning_rate": 2.052823462957142e-05,
      "loss": 0.3779,
      "num_input_tokens_seen": 15009024,
      "step": 22955
    },
    {
      "epoch": 12.033542976939204,
      "grad_norm": 0.3583570718765259,
      "learning_rate": 2.0516985217980953e-05,
      "loss": 0.404,
      "num_input_tokens_seen": 15011680,
      "step": 22960
    },
    {
      "epoch": 12.036163522012579,
      "grad_norm": 0.3544556796550751,
      "learning_rate": 2.050573674418602e-05,
      "loss": 0.4833,
      "num_input_tokens_seen": 15014656,
      "step": 22965
    },
    {
      "epoch": 12.038784067085954,
      "grad_norm": 0.5804100036621094,
      "learning_rate": 2.0494489210539693e-05,
      "loss": 0.3583,
      "num_input_tokens_seen": 15017952,
      "step": 22970
    },
    {
      "epoch": 12.04140461215933,
      "grad_norm": 0.6001423001289368,
      "learning_rate": 2.0483242619394817e-05,
      "loss": 0.5275,
      "num_input_tokens_seen": 15020736,
      "step": 22975
    },
    {
      "epoch": 12.044025157232705,
      "grad_norm": 0.7207878232002258,
      "learning_rate": 2.0471996973104063e-05,
      "loss": 0.5816,
      "num_input_tokens_seen": 15024544,
      "step": 22980
    },
    {
      "epoch": 12.04664570230608,
      "grad_norm": 0.40806350111961365,
      "learning_rate": 2.0460752274019886e-05,
      "loss": 0.4608,
      "num_input_tokens_seen": 15027424,
      "step": 22985
    },
    {
      "epoch": 12.049266247379455,
      "grad_norm": 0.3251676857471466,
      "learning_rate": 2.0449508524494537e-05,
      "loss": 0.3967,
      "num_input_tokens_seen": 15030976,
      "step": 22990
    },
    {
      "epoch": 12.05188679245283,
      "grad_norm": 0.3370383679866791,
      "learning_rate": 2.0438265726880092e-05,
      "loss": 0.4287,
      "num_input_tokens_seen": 15034464,
      "step": 22995
    },
    {
      "epoch": 12.054507337526205,
      "grad_norm": 0.3022483289241791,
      "learning_rate": 2.042702388352841e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 15039104,
      "step": 23000
    },
    {
      "epoch": 12.05712788259958,
      "grad_norm": 0.43995362520217896,
      "learning_rate": 2.041578299679117e-05,
      "loss": 0.3899,
      "num_input_tokens_seen": 15041920,
      "step": 23005
    },
    {
      "epoch": 12.059748427672956,
      "grad_norm": 0.45167750120162964,
      "learning_rate": 2.0404543069019818e-05,
      "loss": 0.5127,
      "num_input_tokens_seen": 15045344,
      "step": 23010
    },
    {
      "epoch": 12.06236897274633,
      "grad_norm": 0.36298346519470215,
      "learning_rate": 2.0393304102565642e-05,
      "loss": 0.4915,
      "num_input_tokens_seen": 15048352,
      "step": 23015
    },
    {
      "epoch": 12.064989517819706,
      "grad_norm": 0.4442290663719177,
      "learning_rate": 2.0382066099779685e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 15051680,
      "step": 23020
    },
    {
      "epoch": 12.067610062893081,
      "grad_norm": 1.3154903650283813,
      "learning_rate": 2.0370829063012808e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 15054304,
      "step": 23025
    },
    {
      "epoch": 12.070230607966456,
      "grad_norm": 0.4084755480289459,
      "learning_rate": 2.035959299461568e-05,
      "loss": 0.383,
      "num_input_tokens_seen": 15056640,
      "step": 23030
    },
    {
      "epoch": 12.072851153039831,
      "grad_norm": 0.46390169858932495,
      "learning_rate": 2.0348357896938762e-05,
      "loss": 0.4952,
      "num_input_tokens_seen": 15059456,
      "step": 23035
    },
    {
      "epoch": 12.075471698113208,
      "grad_norm": 0.35566040873527527,
      "learning_rate": 2.03371237723323e-05,
      "loss": 0.5458,
      "num_input_tokens_seen": 15062720,
      "step": 23040
    },
    {
      "epoch": 12.078092243186584,
      "grad_norm": 0.3247164487838745,
      "learning_rate": 2.032589062314636e-05,
      "loss": 0.3838,
      "num_input_tokens_seen": 15065568,
      "step": 23045
    },
    {
      "epoch": 12.080712788259959,
      "grad_norm": 0.615344226360321,
      "learning_rate": 2.0314658451730767e-05,
      "loss": 0.3737,
      "num_input_tokens_seen": 15068224,
      "step": 23050
    },
    {
      "epoch": 12.083333333333334,
      "grad_norm": 0.36615344882011414,
      "learning_rate": 2.030342726043517e-05,
      "loss": 0.523,
      "num_input_tokens_seen": 15072032,
      "step": 23055
    },
    {
      "epoch": 12.085953878406709,
      "grad_norm": 0.45944157242774963,
      "learning_rate": 2.0292197051609013e-05,
      "loss": 0.3911,
      "num_input_tokens_seen": 15075104,
      "step": 23060
    },
    {
      "epoch": 12.088574423480084,
      "grad_norm": 0.3418426811695099,
      "learning_rate": 2.0280967827601518e-05,
      "loss": 0.5282,
      "num_input_tokens_seen": 15078336,
      "step": 23065
    },
    {
      "epoch": 12.09119496855346,
      "grad_norm": 0.6948022842407227,
      "learning_rate": 2.0269739590761723e-05,
      "loss": 0.3022,
      "num_input_tokens_seen": 15080736,
      "step": 23070
    },
    {
      "epoch": 12.093815513626835,
      "grad_norm": 0.44176656007766724,
      "learning_rate": 2.0258512343438442e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 15083392,
      "step": 23075
    },
    {
      "epoch": 12.09643605870021,
      "grad_norm": 0.5436458587646484,
      "learning_rate": 2.0247286087980276e-05,
      "loss": 0.3276,
      "num_input_tokens_seen": 15086784,
      "step": 23080
    },
    {
      "epoch": 12.099056603773585,
      "grad_norm": 0.5231186747550964,
      "learning_rate": 2.0236060826735644e-05,
      "loss": 0.4695,
      "num_input_tokens_seen": 15089504,
      "step": 23085
    },
    {
      "epoch": 12.10167714884696,
      "grad_norm": 0.3735876977443695,
      "learning_rate": 2.0224836562052732e-05,
      "loss": 0.4889,
      "num_input_tokens_seen": 15092512,
      "step": 23090
    },
    {
      "epoch": 12.104297693920335,
      "grad_norm": 0.3188699781894684,
      "learning_rate": 2.0213613296279533e-05,
      "loss": 0.4697,
      "num_input_tokens_seen": 15095808,
      "step": 23095
    },
    {
      "epoch": 12.10691823899371,
      "grad_norm": 0.41610193252563477,
      "learning_rate": 2.0202391031763844e-05,
      "loss": 0.5514,
      "num_input_tokens_seen": 15098720,
      "step": 23100
    },
    {
      "epoch": 12.109538784067086,
      "grad_norm": 0.556390643119812,
      "learning_rate": 2.0191169770853195e-05,
      "loss": 0.6129,
      "num_input_tokens_seen": 15102176,
      "step": 23105
    },
    {
      "epoch": 12.11215932914046,
      "grad_norm": 1.142152190208435,
      "learning_rate": 2.0179949515894975e-05,
      "loss": 0.4634,
      "num_input_tokens_seen": 15104480,
      "step": 23110
    },
    {
      "epoch": 12.114779874213836,
      "grad_norm": 0.5664827227592468,
      "learning_rate": 2.016873026923633e-05,
      "loss": 0.4401,
      "num_input_tokens_seen": 15107264,
      "step": 23115
    },
    {
      "epoch": 12.117400419287211,
      "grad_norm": 0.3373044729232788,
      "learning_rate": 2.0157512033224187e-05,
      "loss": 0.2803,
      "num_input_tokens_seen": 15111104,
      "step": 23120
    },
    {
      "epoch": 12.120020964360586,
      "grad_norm": 0.4058607220649719,
      "learning_rate": 2.0146294810205287e-05,
      "loss": 0.3579,
      "num_input_tokens_seen": 15114240,
      "step": 23125
    },
    {
      "epoch": 12.122641509433961,
      "grad_norm": 0.46328794956207275,
      "learning_rate": 2.0135078602526146e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 15117600,
      "step": 23130
    },
    {
      "epoch": 12.125262054507338,
      "grad_norm": 0.3195767104625702,
      "learning_rate": 2.0123863412533046e-05,
      "loss": 0.3362,
      "num_input_tokens_seen": 15120416,
      "step": 23135
    },
    {
      "epoch": 12.127882599580714,
      "grad_norm": 0.3635784089565277,
      "learning_rate": 2.0112649242572095e-05,
      "loss": 0.3759,
      "num_input_tokens_seen": 15124064,
      "step": 23140
    },
    {
      "epoch": 12.130503144654089,
      "grad_norm": 0.4077712297439575,
      "learning_rate": 2.0101436094989164e-05,
      "loss": 0.382,
      "num_input_tokens_seen": 15127168,
      "step": 23145
    },
    {
      "epoch": 12.133123689727464,
      "grad_norm": 0.8131842613220215,
      "learning_rate": 2.0090223972129918e-05,
      "loss": 0.3794,
      "num_input_tokens_seen": 15130208,
      "step": 23150
    },
    {
      "epoch": 12.135744234800839,
      "grad_norm": 0.45092493295669556,
      "learning_rate": 2.0079012876339797e-05,
      "loss": 0.5559,
      "num_input_tokens_seen": 15133760,
      "step": 23155
    },
    {
      "epoch": 12.138364779874214,
      "grad_norm": 0.5529482364654541,
      "learning_rate": 2.0067802809964055e-05,
      "loss": 0.5805,
      "num_input_tokens_seen": 15136992,
      "step": 23160
    },
    {
      "epoch": 12.14098532494759,
      "grad_norm": 0.7719618082046509,
      "learning_rate": 2.0056593775347686e-05,
      "loss": 0.4317,
      "num_input_tokens_seen": 15139296,
      "step": 23165
    },
    {
      "epoch": 12.143605870020965,
      "grad_norm": 0.28443020582199097,
      "learning_rate": 2.00453857748355e-05,
      "loss": 0.404,
      "num_input_tokens_seen": 15142656,
      "step": 23170
    },
    {
      "epoch": 12.14622641509434,
      "grad_norm": 0.40282130241394043,
      "learning_rate": 2.0034178810772082e-05,
      "loss": 0.4867,
      "num_input_tokens_seen": 15145824,
      "step": 23175
    },
    {
      "epoch": 12.148846960167715,
      "grad_norm": 0.4244847893714905,
      "learning_rate": 2.002297288550181e-05,
      "loss": 0.3426,
      "num_input_tokens_seen": 15148608,
      "step": 23180
    },
    {
      "epoch": 12.15146750524109,
      "grad_norm": 0.4884498119354248,
      "learning_rate": 2.0011768001368835e-05,
      "loss": 0.4247,
      "num_input_tokens_seen": 15151200,
      "step": 23185
    },
    {
      "epoch": 12.154088050314465,
      "grad_norm": 0.48194876313209534,
      "learning_rate": 2.0000564160717076e-05,
      "loss": 0.4566,
      "num_input_tokens_seen": 15154464,
      "step": 23190
    },
    {
      "epoch": 12.15670859538784,
      "grad_norm": 0.5843838453292847,
      "learning_rate": 1.9989361365890263e-05,
      "loss": 0.5544,
      "num_input_tokens_seen": 15158272,
      "step": 23195
    },
    {
      "epoch": 12.159329140461216,
      "grad_norm": 0.5663546323776245,
      "learning_rate": 1.997815961923188e-05,
      "loss": 0.3558,
      "num_input_tokens_seen": 15161152,
      "step": 23200
    },
    {
      "epoch": 12.16194968553459,
      "grad_norm": 0.5893243551254272,
      "learning_rate": 1.996695892308522e-05,
      "loss": 0.3895,
      "num_input_tokens_seen": 15163616,
      "step": 23205
    },
    {
      "epoch": 12.164570230607966,
      "grad_norm": 0.7136785984039307,
      "learning_rate": 1.9955759279793327e-05,
      "loss": 0.4841,
      "num_input_tokens_seen": 15167264,
      "step": 23210
    },
    {
      "epoch": 12.167190775681341,
      "grad_norm": 0.34711161255836487,
      "learning_rate": 1.9944560691699057e-05,
      "loss": 0.3675,
      "num_input_tokens_seen": 15170816,
      "step": 23215
    },
    {
      "epoch": 12.169811320754716,
      "grad_norm": 0.6816478371620178,
      "learning_rate": 1.9933363161145006e-05,
      "loss": 0.4806,
      "num_input_tokens_seen": 15173280,
      "step": 23220
    },
    {
      "epoch": 12.172431865828091,
      "grad_norm": 0.6995561122894287,
      "learning_rate": 1.9922166690473575e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 15176544,
      "step": 23225
    },
    {
      "epoch": 12.175052410901468,
      "grad_norm": 0.43151214718818665,
      "learning_rate": 1.991097128202695e-05,
      "loss": 0.3558,
      "num_input_tokens_seen": 15179488,
      "step": 23230
    },
    {
      "epoch": 12.177672955974844,
      "grad_norm": 3.7565391063690186,
      "learning_rate": 1.9899776938147065e-05,
      "loss": 0.4346,
      "num_input_tokens_seen": 15182016,
      "step": 23235
    },
    {
      "epoch": 12.180293501048219,
      "grad_norm": 0.3239447772502899,
      "learning_rate": 1.9888583661175665e-05,
      "loss": 0.4839,
      "num_input_tokens_seen": 15185760,
      "step": 23240
    },
    {
      "epoch": 12.182914046121594,
      "grad_norm": 0.4756316840648651,
      "learning_rate": 1.9877391453454253e-05,
      "loss": 0.3971,
      "num_input_tokens_seen": 15188992,
      "step": 23245
    },
    {
      "epoch": 12.185534591194969,
      "grad_norm": 0.37197378277778625,
      "learning_rate": 1.98662003173241e-05,
      "loss": 0.4764,
      "num_input_tokens_seen": 15192768,
      "step": 23250
    },
    {
      "epoch": 12.188155136268344,
      "grad_norm": 0.5409005880355835,
      "learning_rate": 1.9855010255126266e-05,
      "loss": 0.5085,
      "num_input_tokens_seen": 15196256,
      "step": 23255
    },
    {
      "epoch": 12.19077568134172,
      "grad_norm": 0.4088665246963501,
      "learning_rate": 1.9843821269201598e-05,
      "loss": 0.4059,
      "num_input_tokens_seen": 15202848,
      "step": 23260
    },
    {
      "epoch": 12.193396226415095,
      "grad_norm": 1.132548451423645,
      "learning_rate": 1.983263336189069e-05,
      "loss": 0.4395,
      "num_input_tokens_seen": 15206240,
      "step": 23265
    },
    {
      "epoch": 12.19601677148847,
      "grad_norm": 0.6575229167938232,
      "learning_rate": 1.9821446535533937e-05,
      "loss": 0.3236,
      "num_input_tokens_seen": 15208864,
      "step": 23270
    },
    {
      "epoch": 12.198637316561845,
      "grad_norm": 0.7990168333053589,
      "learning_rate": 1.9810260792471492e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 15211968,
      "step": 23275
    },
    {
      "epoch": 12.20125786163522,
      "grad_norm": 0.8392545580863953,
      "learning_rate": 1.979907613504327e-05,
      "loss": 0.3513,
      "num_input_tokens_seen": 15214656,
      "step": 23280
    },
    {
      "epoch": 12.203878406708595,
      "grad_norm": 0.41968557238578796,
      "learning_rate": 1.978789256558899e-05,
      "loss": 0.4776,
      "num_input_tokens_seen": 15218048,
      "step": 23285
    },
    {
      "epoch": 12.20649895178197,
      "grad_norm": 0.4713714122772217,
      "learning_rate": 1.977671008644812e-05,
      "loss": 0.566,
      "num_input_tokens_seen": 15222112,
      "step": 23290
    },
    {
      "epoch": 12.209119496855346,
      "grad_norm": 0.6266520619392395,
      "learning_rate": 1.9765528699959912e-05,
      "loss": 0.4205,
      "num_input_tokens_seen": 15225312,
      "step": 23295
    },
    {
      "epoch": 12.21174004192872,
      "grad_norm": 0.5791075825691223,
      "learning_rate": 1.975434840846339e-05,
      "loss": 0.5671,
      "num_input_tokens_seen": 15228288,
      "step": 23300
    },
    {
      "epoch": 12.214360587002096,
      "grad_norm": 0.40339791774749756,
      "learning_rate": 1.9743169214297326e-05,
      "loss": 0.3373,
      "num_input_tokens_seen": 15232352,
      "step": 23305
    },
    {
      "epoch": 12.216981132075471,
      "grad_norm": 0.28279462456703186,
      "learning_rate": 1.973199111980029e-05,
      "loss": 0.3985,
      "num_input_tokens_seen": 15237984,
      "step": 23310
    },
    {
      "epoch": 12.219601677148846,
      "grad_norm": 0.37472420930862427,
      "learning_rate": 1.9720814127310608e-05,
      "loss": 0.3202,
      "num_input_tokens_seen": 15241184,
      "step": 23315
    },
    {
      "epoch": 12.222222222222221,
      "grad_norm": 0.5871274471282959,
      "learning_rate": 1.9709638239166386e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 15243744,
      "step": 23320
    },
    {
      "epoch": 12.224842767295598,
      "grad_norm": 1.422883152961731,
      "learning_rate": 1.9698463457705485e-05,
      "loss": 0.6552,
      "num_input_tokens_seen": 15246432,
      "step": 23325
    },
    {
      "epoch": 12.227463312368974,
      "grad_norm": 0.5345937609672546,
      "learning_rate": 1.968728978526556e-05,
      "loss": 0.398,
      "num_input_tokens_seen": 15249440,
      "step": 23330
    },
    {
      "epoch": 12.230083857442349,
      "grad_norm": 0.38063186407089233,
      "learning_rate": 1.9676117224183986e-05,
      "loss": 0.4555,
      "num_input_tokens_seen": 15253120,
      "step": 23335
    },
    {
      "epoch": 12.232704402515724,
      "grad_norm": 0.8052985668182373,
      "learning_rate": 1.9664945776797956e-05,
      "loss": 0.3872,
      "num_input_tokens_seen": 15256224,
      "step": 23340
    },
    {
      "epoch": 12.235324947589099,
      "grad_norm": 0.4678226709365845,
      "learning_rate": 1.96537754454444e-05,
      "loss": 0.5884,
      "num_input_tokens_seen": 15258720,
      "step": 23345
    },
    {
      "epoch": 12.237945492662474,
      "grad_norm": 0.4539092183113098,
      "learning_rate": 1.964260623246003e-05,
      "loss": 0.4852,
      "num_input_tokens_seen": 15261760,
      "step": 23350
    },
    {
      "epoch": 12.24056603773585,
      "grad_norm": 0.4144359529018402,
      "learning_rate": 1.9631438140181324e-05,
      "loss": 0.5315,
      "num_input_tokens_seen": 15265472,
      "step": 23355
    },
    {
      "epoch": 12.243186582809225,
      "grad_norm": 0.2508055865764618,
      "learning_rate": 1.9620271170944503e-05,
      "loss": 0.3748,
      "num_input_tokens_seen": 15268896,
      "step": 23360
    },
    {
      "epoch": 12.2458071278826,
      "grad_norm": 1.6004375219345093,
      "learning_rate": 1.960910532708558e-05,
      "loss": 0.507,
      "num_input_tokens_seen": 15272320,
      "step": 23365
    },
    {
      "epoch": 12.248427672955975,
      "grad_norm": 0.7611699104309082,
      "learning_rate": 1.9597940610940315e-05,
      "loss": 0.5248,
      "num_input_tokens_seen": 15275776,
      "step": 23370
    },
    {
      "epoch": 12.25104821802935,
      "grad_norm": 0.5314652919769287,
      "learning_rate": 1.958677702484425e-05,
      "loss": 0.3726,
      "num_input_tokens_seen": 15278912,
      "step": 23375
    },
    {
      "epoch": 12.253668763102725,
      "grad_norm": 0.5688533186912537,
      "learning_rate": 1.957561457113267e-05,
      "loss": 0.4983,
      "num_input_tokens_seen": 15281536,
      "step": 23380
    },
    {
      "epoch": 12.2562893081761,
      "grad_norm": 0.6940938234329224,
      "learning_rate": 1.9564453252140648e-05,
      "loss": 0.4732,
      "num_input_tokens_seen": 15284544,
      "step": 23385
    },
    {
      "epoch": 12.258909853249476,
      "grad_norm": 0.3261057138442993,
      "learning_rate": 1.9553293070202984e-05,
      "loss": 0.3776,
      "num_input_tokens_seen": 15287968,
      "step": 23390
    },
    {
      "epoch": 12.26153039832285,
      "grad_norm": 0.6844142079353333,
      "learning_rate": 1.9542134027654267e-05,
      "loss": 0.3907,
      "num_input_tokens_seen": 15291136,
      "step": 23395
    },
    {
      "epoch": 12.264150943396226,
      "grad_norm": 0.71007239818573,
      "learning_rate": 1.953097612682885e-05,
      "loss": 0.462,
      "num_input_tokens_seen": 15293824,
      "step": 23400
    },
    {
      "epoch": 12.266771488469601,
      "grad_norm": 0.6946272253990173,
      "learning_rate": 1.9519819370060832e-05,
      "loss": 0.3367,
      "num_input_tokens_seen": 15296928,
      "step": 23405
    },
    {
      "epoch": 12.269392033542976,
      "grad_norm": 0.9772297739982605,
      "learning_rate": 1.9508663759684074e-05,
      "loss": 0.379,
      "num_input_tokens_seen": 15299904,
      "step": 23410
    },
    {
      "epoch": 12.272012578616351,
      "grad_norm": 0.44994303584098816,
      "learning_rate": 1.9497509298032224e-05,
      "loss": 0.4685,
      "num_input_tokens_seen": 15302624,
      "step": 23415
    },
    {
      "epoch": 12.274633123689728,
      "grad_norm": 0.4405049681663513,
      "learning_rate": 1.9486355987438647e-05,
      "loss": 0.4466,
      "num_input_tokens_seen": 15305408,
      "step": 23420
    },
    {
      "epoch": 12.277253668763104,
      "grad_norm": 1.1354905366897583,
      "learning_rate": 1.947520383023649e-05,
      "loss": 0.5299,
      "num_input_tokens_seen": 15308288,
      "step": 23425
    },
    {
      "epoch": 12.279874213836479,
      "grad_norm": 0.5110544562339783,
      "learning_rate": 1.9464052828758666e-05,
      "loss": 0.4047,
      "num_input_tokens_seen": 15311200,
      "step": 23430
    },
    {
      "epoch": 12.282494758909854,
      "grad_norm": 0.3025093674659729,
      "learning_rate": 1.945290298533783e-05,
      "loss": 0.5549,
      "num_input_tokens_seen": 15314688,
      "step": 23435
    },
    {
      "epoch": 12.285115303983229,
      "grad_norm": 0.9902418255805969,
      "learning_rate": 1.9441754302306406e-05,
      "loss": 0.434,
      "num_input_tokens_seen": 15317632,
      "step": 23440
    },
    {
      "epoch": 12.287735849056604,
      "grad_norm": 0.457491010427475,
      "learning_rate": 1.9430606781996578e-05,
      "loss": 0.3489,
      "num_input_tokens_seen": 15320288,
      "step": 23445
    },
    {
      "epoch": 12.29035639412998,
      "grad_norm": 0.4026102125644684,
      "learning_rate": 1.941946042674026e-05,
      "loss": 0.4299,
      "num_input_tokens_seen": 15323520,
      "step": 23450
    },
    {
      "epoch": 12.292976939203355,
      "grad_norm": 0.6903731226921082,
      "learning_rate": 1.940831523886916e-05,
      "loss": 0.5289,
      "num_input_tokens_seen": 15327008,
      "step": 23455
    },
    {
      "epoch": 12.29559748427673,
      "grad_norm": 0.37605932354927063,
      "learning_rate": 1.9397171220714714e-05,
      "loss": 0.4295,
      "num_input_tokens_seen": 15329952,
      "step": 23460
    },
    {
      "epoch": 12.298218029350105,
      "grad_norm": 0.48265787959098816,
      "learning_rate": 1.9386028374608134e-05,
      "loss": 0.4803,
      "num_input_tokens_seen": 15333056,
      "step": 23465
    },
    {
      "epoch": 12.30083857442348,
      "grad_norm": 0.41513872146606445,
      "learning_rate": 1.9374886702880372e-05,
      "loss": 0.4208,
      "num_input_tokens_seen": 15336736,
      "step": 23470
    },
    {
      "epoch": 12.303459119496855,
      "grad_norm": 0.5593945384025574,
      "learning_rate": 1.936374620786213e-05,
      "loss": 0.3841,
      "num_input_tokens_seen": 15339712,
      "step": 23475
    },
    {
      "epoch": 12.30607966457023,
      "grad_norm": 0.28739896416664124,
      "learning_rate": 1.9352606891883885e-05,
      "loss": 0.5431,
      "num_input_tokens_seen": 15343232,
      "step": 23480
    },
    {
      "epoch": 12.308700209643606,
      "grad_norm": 0.652130663394928,
      "learning_rate": 1.9341468757275844e-05,
      "loss": 0.4586,
      "num_input_tokens_seen": 15346816,
      "step": 23485
    },
    {
      "epoch": 12.31132075471698,
      "grad_norm": 0.38988929986953735,
      "learning_rate": 1.933033180636798e-05,
      "loss": 0.5344,
      "num_input_tokens_seen": 15349760,
      "step": 23490
    },
    {
      "epoch": 12.313941299790356,
      "grad_norm": 0.3180743455886841,
      "learning_rate": 1.9319196041490027e-05,
      "loss": 0.4438,
      "num_input_tokens_seen": 15353184,
      "step": 23495
    },
    {
      "epoch": 12.316561844863731,
      "grad_norm": 0.31537100672721863,
      "learning_rate": 1.930806146497146e-05,
      "loss": 0.3593,
      "num_input_tokens_seen": 15356512,
      "step": 23500
    },
    {
      "epoch": 12.319182389937106,
      "grad_norm": 0.3860618770122528,
      "learning_rate": 1.9296928079141485e-05,
      "loss": 0.3661,
      "num_input_tokens_seen": 15359104,
      "step": 23505
    },
    {
      "epoch": 12.321802935010481,
      "grad_norm": 0.44187700748443604,
      "learning_rate": 1.9285795886329093e-05,
      "loss": 0.5099,
      "num_input_tokens_seen": 15362528,
      "step": 23510
    },
    {
      "epoch": 12.324423480083858,
      "grad_norm": 0.4906753599643707,
      "learning_rate": 1.927466488886301e-05,
      "loss": 0.5913,
      "num_input_tokens_seen": 15366240,
      "step": 23515
    },
    {
      "epoch": 12.327044025157234,
      "grad_norm": 0.9228895902633667,
      "learning_rate": 1.926353508907172e-05,
      "loss": 0.5093,
      "num_input_tokens_seen": 15368704,
      "step": 23520
    },
    {
      "epoch": 12.329664570230609,
      "grad_norm": 0.44727563858032227,
      "learning_rate": 1.9252406489283437e-05,
      "loss": 0.3857,
      "num_input_tokens_seen": 15373312,
      "step": 23525
    },
    {
      "epoch": 12.332285115303984,
      "grad_norm": 0.6171104907989502,
      "learning_rate": 1.9241279091826155e-05,
      "loss": 0.4708,
      "num_input_tokens_seen": 15377632,
      "step": 23530
    },
    {
      "epoch": 12.334905660377359,
      "grad_norm": 0.4970417618751526,
      "learning_rate": 1.923015289902758e-05,
      "loss": 0.4745,
      "num_input_tokens_seen": 15380864,
      "step": 23535
    },
    {
      "epoch": 12.337526205450734,
      "grad_norm": 0.3116859793663025,
      "learning_rate": 1.9219027913215193e-05,
      "loss": 0.4382,
      "num_input_tokens_seen": 15384480,
      "step": 23540
    },
    {
      "epoch": 12.34014675052411,
      "grad_norm": 0.39202532172203064,
      "learning_rate": 1.9207904136716214e-05,
      "loss": 0.372,
      "num_input_tokens_seen": 15387968,
      "step": 23545
    },
    {
      "epoch": 12.342767295597485,
      "grad_norm": 0.6200741529464722,
      "learning_rate": 1.9196781571857608e-05,
      "loss": 0.3991,
      "num_input_tokens_seen": 15390528,
      "step": 23550
    },
    {
      "epoch": 12.34538784067086,
      "grad_norm": 0.7131155133247375,
      "learning_rate": 1.9185660220966097e-05,
      "loss": 0.4063,
      "num_input_tokens_seen": 15394368,
      "step": 23555
    },
    {
      "epoch": 12.348008385744235,
      "grad_norm": 0.4283427298069,
      "learning_rate": 1.9174540086368135e-05,
      "loss": 0.4161,
      "num_input_tokens_seen": 15397600,
      "step": 23560
    },
    {
      "epoch": 12.35062893081761,
      "grad_norm": 0.49508899450302124,
      "learning_rate": 1.9163421170389918e-05,
      "loss": 0.4009,
      "num_input_tokens_seen": 15400288,
      "step": 23565
    },
    {
      "epoch": 12.353249475890985,
      "grad_norm": 0.43986430764198303,
      "learning_rate": 1.9152303475357402e-05,
      "loss": 0.5323,
      "num_input_tokens_seen": 15404032,
      "step": 23570
    },
    {
      "epoch": 12.35587002096436,
      "grad_norm": 0.5379660725593567,
      "learning_rate": 1.9141187003596295e-05,
      "loss": 0.5911,
      "num_input_tokens_seen": 15407392,
      "step": 23575
    },
    {
      "epoch": 12.358490566037736,
      "grad_norm": 0.665517270565033,
      "learning_rate": 1.913007175743202e-05,
      "loss": 0.3975,
      "num_input_tokens_seen": 15410208,
      "step": 23580
    },
    {
      "epoch": 12.36111111111111,
      "grad_norm": 0.46189093589782715,
      "learning_rate": 1.9118957739189773e-05,
      "loss": 0.5238,
      "num_input_tokens_seen": 15412992,
      "step": 23585
    },
    {
      "epoch": 12.363731656184486,
      "grad_norm": 0.34737923741340637,
      "learning_rate": 1.910784495119447e-05,
      "loss": 0.4912,
      "num_input_tokens_seen": 15417056,
      "step": 23590
    },
    {
      "epoch": 12.366352201257861,
      "grad_norm": 0.82114577293396,
      "learning_rate": 1.9096733395770778e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 15420032,
      "step": 23595
    },
    {
      "epoch": 12.368972746331236,
      "grad_norm": 0.2880573272705078,
      "learning_rate": 1.9085623075243114e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 15423776,
      "step": 23600
    },
    {
      "epoch": 12.371593291404611,
      "grad_norm": 0.4613072872161865,
      "learning_rate": 1.9074513991935623e-05,
      "loss": 0.485,
      "num_input_tokens_seen": 15426880,
      "step": 23605
    },
    {
      "epoch": 12.374213836477987,
      "grad_norm": 0.4177510738372803,
      "learning_rate": 1.906340614817221e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 15430688,
      "step": 23610
    },
    {
      "epoch": 12.376834381551364,
      "grad_norm": 0.6385961771011353,
      "learning_rate": 1.9052299546276506e-05,
      "loss": 0.3743,
      "num_input_tokens_seen": 15433728,
      "step": 23615
    },
    {
      "epoch": 12.379454926624739,
      "grad_norm": 0.48441341519355774,
      "learning_rate": 1.904119418857187e-05,
      "loss": 0.4844,
      "num_input_tokens_seen": 15437568,
      "step": 23620
    },
    {
      "epoch": 12.382075471698114,
      "grad_norm": 0.48930594325065613,
      "learning_rate": 1.903009007738144e-05,
      "loss": 0.4802,
      "num_input_tokens_seen": 15440384,
      "step": 23625
    },
    {
      "epoch": 12.384696016771489,
      "grad_norm": 0.3772335350513458,
      "learning_rate": 1.9018987215028045e-05,
      "loss": 0.4037,
      "num_input_tokens_seen": 15444032,
      "step": 23630
    },
    {
      "epoch": 12.387316561844864,
      "grad_norm": 0.7962745428085327,
      "learning_rate": 1.9007885603834298e-05,
      "loss": 0.3969,
      "num_input_tokens_seen": 15446400,
      "step": 23635
    },
    {
      "epoch": 12.38993710691824,
      "grad_norm": 0.482650488615036,
      "learning_rate": 1.8996785246122526e-05,
      "loss": 0.424,
      "num_input_tokens_seen": 15450880,
      "step": 23640
    },
    {
      "epoch": 12.392557651991615,
      "grad_norm": 0.3907569348812103,
      "learning_rate": 1.898568614421478e-05,
      "loss": 0.4843,
      "num_input_tokens_seen": 15453856,
      "step": 23645
    },
    {
      "epoch": 12.39517819706499,
      "grad_norm": 0.7137621641159058,
      "learning_rate": 1.897458830043288e-05,
      "loss": 0.3736,
      "num_input_tokens_seen": 15456320,
      "step": 23650
    },
    {
      "epoch": 12.397798742138365,
      "grad_norm": 0.5088328123092651,
      "learning_rate": 1.8963491717098372e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 15459616,
      "step": 23655
    },
    {
      "epoch": 12.40041928721174,
      "grad_norm": 0.5275649428367615,
      "learning_rate": 1.8952396396532522e-05,
      "loss": 0.4928,
      "num_input_tokens_seen": 15463520,
      "step": 23660
    },
    {
      "epoch": 12.403039832285115,
      "grad_norm": 0.5389566421508789,
      "learning_rate": 1.8941302341056356e-05,
      "loss": 0.4092,
      "num_input_tokens_seen": 15466976,
      "step": 23665
    },
    {
      "epoch": 12.40566037735849,
      "grad_norm": 0.5798121690750122,
      "learning_rate": 1.8930209552990625e-05,
      "loss": 0.6073,
      "num_input_tokens_seen": 15470784,
      "step": 23670
    },
    {
      "epoch": 12.408280922431866,
      "grad_norm": 0.4071049094200134,
      "learning_rate": 1.8919118034655796e-05,
      "loss": 0.3386,
      "num_input_tokens_seen": 15474816,
      "step": 23675
    },
    {
      "epoch": 12.41090146750524,
      "grad_norm": 0.9530402421951294,
      "learning_rate": 1.8908027788372107e-05,
      "loss": 0.4802,
      "num_input_tokens_seen": 15478016,
      "step": 23680
    },
    {
      "epoch": 12.413522012578616,
      "grad_norm": 0.3442113995552063,
      "learning_rate": 1.88969388164595e-05,
      "loss": 0.3547,
      "num_input_tokens_seen": 15481536,
      "step": 23685
    },
    {
      "epoch": 12.416142557651991,
      "grad_norm": 0.3369910717010498,
      "learning_rate": 1.8885851121237666e-05,
      "loss": 0.3499,
      "num_input_tokens_seen": 15485504,
      "step": 23690
    },
    {
      "epoch": 12.418763102725366,
      "grad_norm": 0.5437930226325989,
      "learning_rate": 1.8874764705026023e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 15489120,
      "step": 23695
    },
    {
      "epoch": 12.421383647798741,
      "grad_norm": 0.704826295375824,
      "learning_rate": 1.8863679570143728e-05,
      "loss": 0.5445,
      "num_input_tokens_seen": 15491840,
      "step": 23700
    },
    {
      "epoch": 12.424004192872117,
      "grad_norm": 0.48940443992614746,
      "learning_rate": 1.885259571890966e-05,
      "loss": 0.3623,
      "num_input_tokens_seen": 15494880,
      "step": 23705
    },
    {
      "epoch": 12.426624737945493,
      "grad_norm": 0.5691293478012085,
      "learning_rate": 1.8841513153642427e-05,
      "loss": 0.3807,
      "num_input_tokens_seen": 15497312,
      "step": 23710
    },
    {
      "epoch": 12.429245283018869,
      "grad_norm": 0.43647298216819763,
      "learning_rate": 1.8830431876660388e-05,
      "loss": 0.364,
      "num_input_tokens_seen": 15500416,
      "step": 23715
    },
    {
      "epoch": 12.431865828092244,
      "grad_norm": 0.9603444933891296,
      "learning_rate": 1.8819351890281607e-05,
      "loss": 0.5127,
      "num_input_tokens_seen": 15503520,
      "step": 23720
    },
    {
      "epoch": 12.434486373165619,
      "grad_norm": 0.49602797627449036,
      "learning_rate": 1.880827319682391e-05,
      "loss": 0.3247,
      "num_input_tokens_seen": 15506656,
      "step": 23725
    },
    {
      "epoch": 12.437106918238994,
      "grad_norm": 0.4617442488670349,
      "learning_rate": 1.879719579860481e-05,
      "loss": 0.4805,
      "num_input_tokens_seen": 15509792,
      "step": 23730
    },
    {
      "epoch": 12.43972746331237,
      "grad_norm": 0.6334390044212341,
      "learning_rate": 1.878611969794159e-05,
      "loss": 0.4084,
      "num_input_tokens_seen": 15512768,
      "step": 23735
    },
    {
      "epoch": 12.442348008385745,
      "grad_norm": 0.519347608089447,
      "learning_rate": 1.8775044897151232e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 15515712,
      "step": 23740
    },
    {
      "epoch": 12.44496855345912,
      "grad_norm": 0.5308158993721008,
      "learning_rate": 1.876397139855047e-05,
      "loss": 0.546,
      "num_input_tokens_seen": 15518528,
      "step": 23745
    },
    {
      "epoch": 12.447589098532495,
      "grad_norm": 0.5026542544364929,
      "learning_rate": 1.875289920445574e-05,
      "loss": 0.5309,
      "num_input_tokens_seen": 15521472,
      "step": 23750
    },
    {
      "epoch": 12.45020964360587,
      "grad_norm": 1.8068084716796875,
      "learning_rate": 1.8741828317183234e-05,
      "loss": 0.324,
      "num_input_tokens_seen": 15524320,
      "step": 23755
    },
    {
      "epoch": 12.452830188679245,
      "grad_norm": 0.48404672741889954,
      "learning_rate": 1.873075873904885e-05,
      "loss": 0.4715,
      "num_input_tokens_seen": 15527168,
      "step": 23760
    },
    {
      "epoch": 12.45545073375262,
      "grad_norm": 0.5273168683052063,
      "learning_rate": 1.87196904723682e-05,
      "loss": 0.3481,
      "num_input_tokens_seen": 15532064,
      "step": 23765
    },
    {
      "epoch": 12.458071278825996,
      "grad_norm": 0.4958290755748749,
      "learning_rate": 1.8708623519456665e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 15534944,
      "step": 23770
    },
    {
      "epoch": 12.46069182389937,
      "grad_norm": 0.4034421443939209,
      "learning_rate": 1.8697557882629308e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 15538368,
      "step": 23775
    },
    {
      "epoch": 12.463312368972746,
      "grad_norm": 0.37346744537353516,
      "learning_rate": 1.868649356420094e-05,
      "loss": 0.4053,
      "num_input_tokens_seen": 15541280,
      "step": 23780
    },
    {
      "epoch": 12.465932914046121,
      "grad_norm": 0.4952343702316284,
      "learning_rate": 1.86754305664861e-05,
      "loss": 0.3809,
      "num_input_tokens_seen": 15544736,
      "step": 23785
    },
    {
      "epoch": 12.468553459119496,
      "grad_norm": 0.509320855140686,
      "learning_rate": 1.866436889179902e-05,
      "loss": 0.5065,
      "num_input_tokens_seen": 15549568,
      "step": 23790
    },
    {
      "epoch": 12.471174004192871,
      "grad_norm": 0.5991253852844238,
      "learning_rate": 1.865330854245369e-05,
      "loss": 0.3995,
      "num_input_tokens_seen": 15553184,
      "step": 23795
    },
    {
      "epoch": 12.473794549266247,
      "grad_norm": 0.3344303369522095,
      "learning_rate": 1.8642249520763804e-05,
      "loss": 0.4142,
      "num_input_tokens_seen": 15557120,
      "step": 23800
    },
    {
      "epoch": 12.476415094339623,
      "grad_norm": 0.6766633987426758,
      "learning_rate": 1.863119182904278e-05,
      "loss": 0.4065,
      "num_input_tokens_seen": 15561632,
      "step": 23805
    },
    {
      "epoch": 12.479035639412999,
      "grad_norm": 0.4895882308483124,
      "learning_rate": 1.8620135469603777e-05,
      "loss": 0.2942,
      "num_input_tokens_seen": 15568736,
      "step": 23810
    },
    {
      "epoch": 12.481656184486374,
      "grad_norm": 0.41805630922317505,
      "learning_rate": 1.8609080444759653e-05,
      "loss": 0.505,
      "num_input_tokens_seen": 15572352,
      "step": 23815
    },
    {
      "epoch": 12.484276729559749,
      "grad_norm": 0.6664091944694519,
      "learning_rate": 1.8598026756822978e-05,
      "loss": 0.3969,
      "num_input_tokens_seen": 15574688,
      "step": 23820
    },
    {
      "epoch": 12.486897274633124,
      "grad_norm": 0.322866827249527,
      "learning_rate": 1.8586974408106074e-05,
      "loss": 0.4896,
      "num_input_tokens_seen": 15577376,
      "step": 23825
    },
    {
      "epoch": 12.4895178197065,
      "grad_norm": 0.47060340642929077,
      "learning_rate": 1.8575923400920954e-05,
      "loss": 0.4708,
      "num_input_tokens_seen": 15581280,
      "step": 23830
    },
    {
      "epoch": 12.492138364779874,
      "grad_norm": 0.41497236490249634,
      "learning_rate": 1.856487373757937e-05,
      "loss": 0.4857,
      "num_input_tokens_seen": 15584544,
      "step": 23835
    },
    {
      "epoch": 12.49475890985325,
      "grad_norm": 0.5355193614959717,
      "learning_rate": 1.8553825420392792e-05,
      "loss": 0.428,
      "num_input_tokens_seen": 15587520,
      "step": 23840
    },
    {
      "epoch": 12.497379454926625,
      "grad_norm": 0.856330931186676,
      "learning_rate": 1.8542778451672385e-05,
      "loss": 0.7238,
      "num_input_tokens_seen": 15590624,
      "step": 23845
    },
    {
      "epoch": 12.5,
      "grad_norm": 0.4162449240684509,
      "learning_rate": 1.8531732833729063e-05,
      "loss": 0.4908,
      "num_input_tokens_seen": 15594560,
      "step": 23850
    },
    {
      "epoch": 12.502620545073375,
      "grad_norm": 0.3327876627445221,
      "learning_rate": 1.852068856887343e-05,
      "loss": 0.5767,
      "num_input_tokens_seen": 15598048,
      "step": 23855
    },
    {
      "epoch": 12.50524109014675,
      "grad_norm": 0.40419867634773254,
      "learning_rate": 1.8509645659415837e-05,
      "loss": 0.4919,
      "num_input_tokens_seen": 15602048,
      "step": 23860
    },
    {
      "epoch": 12.507861635220126,
      "grad_norm": 0.6111318469047546,
      "learning_rate": 1.849860410766632e-05,
      "loss": 0.5481,
      "num_input_tokens_seen": 15605152,
      "step": 23865
    },
    {
      "epoch": 12.5104821802935,
      "grad_norm": 0.674619197845459,
      "learning_rate": 1.8487563915934664e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 15607808,
      "step": 23870
    },
    {
      "epoch": 12.513102725366876,
      "grad_norm": 0.13937821984291077,
      "learning_rate": 1.847652508653033e-05,
      "loss": 0.4594,
      "num_input_tokens_seen": 15613344,
      "step": 23875
    },
    {
      "epoch": 12.515723270440251,
      "grad_norm": 0.3536813259124756,
      "learning_rate": 1.846548762176252e-05,
      "loss": 0.5443,
      "num_input_tokens_seen": 15616128,
      "step": 23880
    },
    {
      "epoch": 12.518343815513626,
      "grad_norm": 0.36140909790992737,
      "learning_rate": 1.845445152394015e-05,
      "loss": 0.4998,
      "num_input_tokens_seen": 15619328,
      "step": 23885
    },
    {
      "epoch": 12.520964360587001,
      "grad_norm": 0.27794501185417175,
      "learning_rate": 1.8443416795371853e-05,
      "loss": 0.4818,
      "num_input_tokens_seen": 15622624,
      "step": 23890
    },
    {
      "epoch": 12.523584905660378,
      "grad_norm": 0.28968533873558044,
      "learning_rate": 1.8432383438365957e-05,
      "loss": 0.3643,
      "num_input_tokens_seen": 15626560,
      "step": 23895
    },
    {
      "epoch": 12.526205450733752,
      "grad_norm": 0.4656193256378174,
      "learning_rate": 1.8421351455230528e-05,
      "loss": 0.5358,
      "num_input_tokens_seen": 15629888,
      "step": 23900
    },
    {
      "epoch": 12.528825995807129,
      "grad_norm": 0.6626493334770203,
      "learning_rate": 1.8410320848273315e-05,
      "loss": 0.6604,
      "num_input_tokens_seen": 15633312,
      "step": 23905
    },
    {
      "epoch": 12.531446540880504,
      "grad_norm": 0.4564690887928009,
      "learning_rate": 1.8399291619801794e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 15636096,
      "step": 23910
    },
    {
      "epoch": 12.534067085953879,
      "grad_norm": 1.3816932439804077,
      "learning_rate": 1.838826377212317e-05,
      "loss": 0.4463,
      "num_input_tokens_seen": 15638496,
      "step": 23915
    },
    {
      "epoch": 12.536687631027254,
      "grad_norm": 0.30573248863220215,
      "learning_rate": 1.8377237307544326e-05,
      "loss": 0.3611,
      "num_input_tokens_seen": 15642848,
      "step": 23920
    },
    {
      "epoch": 12.53930817610063,
      "grad_norm": 0.35841765999794006,
      "learning_rate": 1.8366212228371894e-05,
      "loss": 0.4053,
      "num_input_tokens_seen": 15646080,
      "step": 23925
    },
    {
      "epoch": 12.541928721174004,
      "grad_norm": 0.3925415277481079,
      "learning_rate": 1.8355188536912173e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 15648928,
      "step": 23930
    },
    {
      "epoch": 12.54454926624738,
      "grad_norm": 0.4605412483215332,
      "learning_rate": 1.83441662354712e-05,
      "loss": 0.471,
      "num_input_tokens_seen": 15652608,
      "step": 23935
    },
    {
      "epoch": 12.547169811320755,
      "grad_norm": 0.2835805416107178,
      "learning_rate": 1.8333145326354722e-05,
      "loss": 0.3753,
      "num_input_tokens_seen": 15656576,
      "step": 23940
    },
    {
      "epoch": 12.54979035639413,
      "grad_norm": 0.4626876413822174,
      "learning_rate": 1.8322125811868175e-05,
      "loss": 0.4932,
      "num_input_tokens_seen": 15659264,
      "step": 23945
    },
    {
      "epoch": 12.552410901467505,
      "grad_norm": 0.7783025503158569,
      "learning_rate": 1.831110769431673e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 15662272,
      "step": 23950
    },
    {
      "epoch": 12.55503144654088,
      "grad_norm": 0.5221035480499268,
      "learning_rate": 1.8300090976005247e-05,
      "loss": 0.3706,
      "num_input_tokens_seen": 15665184,
      "step": 23955
    },
    {
      "epoch": 12.557651991614255,
      "grad_norm": 0.5041978359222412,
      "learning_rate": 1.8289075659238285e-05,
      "loss": 0.3574,
      "num_input_tokens_seen": 15668832,
      "step": 23960
    },
    {
      "epoch": 12.56027253668763,
      "grad_norm": 0.3406658470630646,
      "learning_rate": 1.8278061746320134e-05,
      "loss": 0.5457,
      "num_input_tokens_seen": 15672160,
      "step": 23965
    },
    {
      "epoch": 12.562893081761006,
      "grad_norm": 0.2805364429950714,
      "learning_rate": 1.8267049239554786e-05,
      "loss": 0.3876,
      "num_input_tokens_seen": 15675296,
      "step": 23970
    },
    {
      "epoch": 12.565513626834381,
      "grad_norm": 0.5221624970436096,
      "learning_rate": 1.8256038141245912e-05,
      "loss": 0.4274,
      "num_input_tokens_seen": 15678784,
      "step": 23975
    },
    {
      "epoch": 12.568134171907756,
      "grad_norm": 0.47969070076942444,
      "learning_rate": 1.824502845369693e-05,
      "loss": 0.4824,
      "num_input_tokens_seen": 15680992,
      "step": 23980
    },
    {
      "epoch": 12.570754716981131,
      "grad_norm": 0.5456130504608154,
      "learning_rate": 1.8234020179210938e-05,
      "loss": 0.433,
      "num_input_tokens_seen": 15683872,
      "step": 23985
    },
    {
      "epoch": 12.573375262054507,
      "grad_norm": 0.5398575067520142,
      "learning_rate": 1.8223013320090715e-05,
      "loss": 0.4666,
      "num_input_tokens_seen": 15686432,
      "step": 23990
    },
    {
      "epoch": 12.575995807127882,
      "grad_norm": 0.41472873091697693,
      "learning_rate": 1.82120078786388e-05,
      "loss": 0.3632,
      "num_input_tokens_seen": 15691296,
      "step": 23995
    },
    {
      "epoch": 12.578616352201259,
      "grad_norm": 0.41508692502975464,
      "learning_rate": 1.820100385715739e-05,
      "loss": 0.5752,
      "num_input_tokens_seen": 15695712,
      "step": 24000
    },
    {
      "epoch": 12.581236897274634,
      "grad_norm": 0.3634118139743805,
      "learning_rate": 1.8190001257948414e-05,
      "loss": 0.3616,
      "num_input_tokens_seen": 15699264,
      "step": 24005
    },
    {
      "epoch": 12.583857442348009,
      "grad_norm": 0.4766310453414917,
      "learning_rate": 1.8179000083313483e-05,
      "loss": 0.4596,
      "num_input_tokens_seen": 15701952,
      "step": 24010
    },
    {
      "epoch": 12.586477987421384,
      "grad_norm": 0.33980655670166016,
      "learning_rate": 1.8168000335553907e-05,
      "loss": 0.3751,
      "num_input_tokens_seen": 15704768,
      "step": 24015
    },
    {
      "epoch": 12.58909853249476,
      "grad_norm": 0.57769376039505,
      "learning_rate": 1.8157002016970723e-05,
      "loss": 0.5378,
      "num_input_tokens_seen": 15708352,
      "step": 24020
    },
    {
      "epoch": 12.591719077568134,
      "grad_norm": 0.4472511112689972,
      "learning_rate": 1.8146005129864644e-05,
      "loss": 0.4692,
      "num_input_tokens_seen": 15711424,
      "step": 24025
    },
    {
      "epoch": 12.59433962264151,
      "grad_norm": 0.4653109014034271,
      "learning_rate": 1.81350096765361e-05,
      "loss": 0.468,
      "num_input_tokens_seen": 15714848,
      "step": 24030
    },
    {
      "epoch": 12.596960167714885,
      "grad_norm": 0.9539782404899597,
      "learning_rate": 1.812401565928521e-05,
      "loss": 0.4825,
      "num_input_tokens_seen": 15717664,
      "step": 24035
    },
    {
      "epoch": 12.59958071278826,
      "grad_norm": 0.47683238983154297,
      "learning_rate": 1.8113023080411806e-05,
      "loss": 0.4668,
      "num_input_tokens_seen": 15720448,
      "step": 24040
    },
    {
      "epoch": 12.602201257861635,
      "grad_norm": 0.7568230032920837,
      "learning_rate": 1.810203194221539e-05,
      "loss": 0.4235,
      "num_input_tokens_seen": 15723552,
      "step": 24045
    },
    {
      "epoch": 12.60482180293501,
      "grad_norm": 0.5784202218055725,
      "learning_rate": 1.8091042246995206e-05,
      "loss": 0.3843,
      "num_input_tokens_seen": 15727616,
      "step": 24050
    },
    {
      "epoch": 12.607442348008385,
      "grad_norm": 0.5488072037696838,
      "learning_rate": 1.8080053997050154e-05,
      "loss": 0.4615,
      "num_input_tokens_seen": 15730240,
      "step": 24055
    },
    {
      "epoch": 12.61006289308176,
      "grad_norm": 0.4430646598339081,
      "learning_rate": 1.8069067194678867e-05,
      "loss": 0.5152,
      "num_input_tokens_seen": 15734048,
      "step": 24060
    },
    {
      "epoch": 12.612683438155136,
      "grad_norm": 0.5762289762496948,
      "learning_rate": 1.8058081842179642e-05,
      "loss": 0.4491,
      "num_input_tokens_seen": 15736896,
      "step": 24065
    },
    {
      "epoch": 12.615303983228511,
      "grad_norm": 0.588966965675354,
      "learning_rate": 1.804709794185051e-05,
      "loss": 0.4981,
      "num_input_tokens_seen": 15739872,
      "step": 24070
    },
    {
      "epoch": 12.617924528301886,
      "grad_norm": 0.22696341574192047,
      "learning_rate": 1.8036115495989164e-05,
      "loss": 0.4569,
      "num_input_tokens_seen": 15743200,
      "step": 24075
    },
    {
      "epoch": 12.620545073375261,
      "grad_norm": 0.37459665536880493,
      "learning_rate": 1.8025134506892997e-05,
      "loss": 0.4288,
      "num_input_tokens_seen": 15748224,
      "step": 24080
    },
    {
      "epoch": 12.623165618448636,
      "grad_norm": 0.6597583889961243,
      "learning_rate": 1.8014154976859126e-05,
      "loss": 0.4049,
      "num_input_tokens_seen": 15751168,
      "step": 24085
    },
    {
      "epoch": 12.625786163522012,
      "grad_norm": 0.341757208108902,
      "learning_rate": 1.800317690818433e-05,
      "loss": 0.3938,
      "num_input_tokens_seen": 15754592,
      "step": 24090
    },
    {
      "epoch": 12.628406708595389,
      "grad_norm": 0.5852591395378113,
      "learning_rate": 1.799220030316511e-05,
      "loss": 0.4278,
      "num_input_tokens_seen": 15758368,
      "step": 24095
    },
    {
      "epoch": 12.631027253668764,
      "grad_norm": 0.44164785742759705,
      "learning_rate": 1.7981225164097633e-05,
      "loss": 0.4175,
      "num_input_tokens_seen": 15761504,
      "step": 24100
    },
    {
      "epoch": 12.633647798742139,
      "grad_norm": 0.4313649535179138,
      "learning_rate": 1.7970251493277774e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 15764224,
      "step": 24105
    },
    {
      "epoch": 12.636268343815514,
      "grad_norm": 0.745259702205658,
      "learning_rate": 1.7959279293001102e-05,
      "loss": 0.4463,
      "num_input_tokens_seen": 15767072,
      "step": 24110
    },
    {
      "epoch": 12.63888888888889,
      "grad_norm": 0.9143010377883911,
      "learning_rate": 1.7948308565562876e-05,
      "loss": 0.6675,
      "num_input_tokens_seen": 15769920,
      "step": 24115
    },
    {
      "epoch": 12.641509433962264,
      "grad_norm": 0.9304695129394531,
      "learning_rate": 1.793733931325805e-05,
      "loss": 0.4976,
      "num_input_tokens_seen": 15773472,
      "step": 24120
    },
    {
      "epoch": 12.64412997903564,
      "grad_norm": 0.5368394255638123,
      "learning_rate": 1.792637153838127e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 15776064,
      "step": 24125
    },
    {
      "epoch": 12.646750524109015,
      "grad_norm": 0.41290563344955444,
      "learning_rate": 1.7915405243226864e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 15779552,
      "step": 24130
    },
    {
      "epoch": 12.64937106918239,
      "grad_norm": 0.3482309579849243,
      "learning_rate": 1.7904440430088847e-05,
      "loss": 0.2521,
      "num_input_tokens_seen": 15782272,
      "step": 24135
    },
    {
      "epoch": 12.651991614255765,
      "grad_norm": 0.9552116990089417,
      "learning_rate": 1.7893477101260947e-05,
      "loss": 0.5675,
      "num_input_tokens_seen": 15785024,
      "step": 24140
    },
    {
      "epoch": 12.65461215932914,
      "grad_norm": 0.4583004415035248,
      "learning_rate": 1.7882515259036558e-05,
      "loss": 0.6592,
      "num_input_tokens_seen": 15788128,
      "step": 24145
    },
    {
      "epoch": 12.657232704402515,
      "grad_norm": 0.6227560043334961,
      "learning_rate": 1.787155490570878e-05,
      "loss": 0.404,
      "num_input_tokens_seen": 15790944,
      "step": 24150
    },
    {
      "epoch": 12.65985324947589,
      "grad_norm": 0.4424077272415161,
      "learning_rate": 1.7860596043570398e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 15793920,
      "step": 24155
    },
    {
      "epoch": 12.662473794549266,
      "grad_norm": 0.3440626561641693,
      "learning_rate": 1.7849638674913863e-05,
      "loss": 0.2668,
      "num_input_tokens_seen": 15799552,
      "step": 24160
    },
    {
      "epoch": 12.665094339622641,
      "grad_norm": 0.3933180570602417,
      "learning_rate": 1.7838682802031344e-05,
      "loss": 0.4271,
      "num_input_tokens_seen": 15802656,
      "step": 24165
    },
    {
      "epoch": 12.667714884696016,
      "grad_norm": 0.6311271786689758,
      "learning_rate": 1.782772842721468e-05,
      "loss": 0.486,
      "num_input_tokens_seen": 15806112,
      "step": 24170
    },
    {
      "epoch": 12.670335429769391,
      "grad_norm": 0.5889833569526672,
      "learning_rate": 1.7816775552755406e-05,
      "loss": 0.4441,
      "num_input_tokens_seen": 15808928,
      "step": 24175
    },
    {
      "epoch": 12.672955974842766,
      "grad_norm": 0.6688258647918701,
      "learning_rate": 1.780582418094473e-05,
      "loss": 0.4531,
      "num_input_tokens_seen": 15812320,
      "step": 24180
    },
    {
      "epoch": 12.675576519916142,
      "grad_norm": 0.27986136078834534,
      "learning_rate": 1.7794874314073573e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 15816608,
      "step": 24185
    },
    {
      "epoch": 12.678197064989519,
      "grad_norm": 0.45223894715309143,
      "learning_rate": 1.7783925954432503e-05,
      "loss": 0.4432,
      "num_input_tokens_seen": 15819680,
      "step": 24190
    },
    {
      "epoch": 12.680817610062894,
      "grad_norm": 0.554951548576355,
      "learning_rate": 1.777297910431179e-05,
      "loss": 0.4934,
      "num_input_tokens_seen": 15822848,
      "step": 24195
    },
    {
      "epoch": 12.683438155136269,
      "grad_norm": 0.4171094000339508,
      "learning_rate": 1.7762033766001402e-05,
      "loss": 0.5439,
      "num_input_tokens_seen": 15825312,
      "step": 24200
    },
    {
      "epoch": 12.686058700209644,
      "grad_norm": 0.2423667013645172,
      "learning_rate": 1.7751089941790977e-05,
      "loss": 0.3303,
      "num_input_tokens_seen": 15828672,
      "step": 24205
    },
    {
      "epoch": 12.68867924528302,
      "grad_norm": 0.4894350469112396,
      "learning_rate": 1.7740147633969844e-05,
      "loss": 0.4768,
      "num_input_tokens_seen": 15831648,
      "step": 24210
    },
    {
      "epoch": 12.691299790356394,
      "grad_norm": 0.3988124430179596,
      "learning_rate": 1.772920684482699e-05,
      "loss": 0.4821,
      "num_input_tokens_seen": 15835456,
      "step": 24215
    },
    {
      "epoch": 12.69392033542977,
      "grad_norm": 0.4404961168766022,
      "learning_rate": 1.771826757665112e-05,
      "loss": 0.3572,
      "num_input_tokens_seen": 15839008,
      "step": 24220
    },
    {
      "epoch": 12.696540880503145,
      "grad_norm": 0.7231921553611755,
      "learning_rate": 1.770732983173059e-05,
      "loss": 0.5398,
      "num_input_tokens_seen": 15842048,
      "step": 24225
    },
    {
      "epoch": 12.69916142557652,
      "grad_norm": 0.7015446424484253,
      "learning_rate": 1.7696393612353467e-05,
      "loss": 0.4694,
      "num_input_tokens_seen": 15844992,
      "step": 24230
    },
    {
      "epoch": 12.701781970649895,
      "grad_norm": 0.5303786396980286,
      "learning_rate": 1.7685458920807474e-05,
      "loss": 0.3038,
      "num_input_tokens_seen": 15847008,
      "step": 24235
    },
    {
      "epoch": 12.70440251572327,
      "grad_norm": 0.43350401520729065,
      "learning_rate": 1.7674525759380033e-05,
      "loss": 0.5704,
      "num_input_tokens_seen": 15849760,
      "step": 24240
    },
    {
      "epoch": 12.707023060796645,
      "grad_norm": 0.4006085693836212,
      "learning_rate": 1.766359413035823e-05,
      "loss": 0.3805,
      "num_input_tokens_seen": 15852608,
      "step": 24245
    },
    {
      "epoch": 12.70964360587002,
      "grad_norm": 0.6699107885360718,
      "learning_rate": 1.7652664036028827e-05,
      "loss": 0.4132,
      "num_input_tokens_seen": 15855904,
      "step": 24250
    },
    {
      "epoch": 12.712264150943396,
      "grad_norm": 0.43153116106987,
      "learning_rate": 1.7641735478678288e-05,
      "loss": 0.3932,
      "num_input_tokens_seen": 15860224,
      "step": 24255
    },
    {
      "epoch": 12.714884696016771,
      "grad_norm": 0.5592233538627625,
      "learning_rate": 1.763080846059274e-05,
      "loss": 0.3981,
      "num_input_tokens_seen": 15863136,
      "step": 24260
    },
    {
      "epoch": 12.717505241090146,
      "grad_norm": 0.5260426998138428,
      "learning_rate": 1.7619882984057987e-05,
      "loss": 0.4902,
      "num_input_tokens_seen": 15866400,
      "step": 24265
    },
    {
      "epoch": 12.720125786163521,
      "grad_norm": 0.5899391770362854,
      "learning_rate": 1.760895905135953e-05,
      "loss": 0.4618,
      "num_input_tokens_seen": 15868896,
      "step": 24270
    },
    {
      "epoch": 12.722746331236896,
      "grad_norm": 0.8001201748847961,
      "learning_rate": 1.7598036664782508e-05,
      "loss": 0.5339,
      "num_input_tokens_seen": 15871232,
      "step": 24275
    },
    {
      "epoch": 12.725366876310272,
      "grad_norm": 0.5239823460578918,
      "learning_rate": 1.7587115826611766e-05,
      "loss": 0.3405,
      "num_input_tokens_seen": 15876800,
      "step": 24280
    },
    {
      "epoch": 12.727987421383649,
      "grad_norm": 0.6571948528289795,
      "learning_rate": 1.7576196539131828e-05,
      "loss": 0.2616,
      "num_input_tokens_seen": 15879520,
      "step": 24285
    },
    {
      "epoch": 12.730607966457024,
      "grad_norm": 0.40774229168891907,
      "learning_rate": 1.7565278804626872e-05,
      "loss": 0.4434,
      "num_input_tokens_seen": 15882432,
      "step": 24290
    },
    {
      "epoch": 12.733228511530399,
      "grad_norm": 0.26810482144355774,
      "learning_rate": 1.755436262538078e-05,
      "loss": 0.4949,
      "num_input_tokens_seen": 15887648,
      "step": 24295
    },
    {
      "epoch": 12.735849056603774,
      "grad_norm": 0.7098177075386047,
      "learning_rate": 1.7543448003677082e-05,
      "loss": 0.3979,
      "num_input_tokens_seen": 15892704,
      "step": 24300
    },
    {
      "epoch": 12.73846960167715,
      "grad_norm": 0.32399672269821167,
      "learning_rate": 1.7532534941798985e-05,
      "loss": 0.4188,
      "num_input_tokens_seen": 15895840,
      "step": 24305
    },
    {
      "epoch": 12.741090146750524,
      "grad_norm": 0.5267035961151123,
      "learning_rate": 1.752162344202939e-05,
      "loss": 0.5777,
      "num_input_tokens_seen": 15899488,
      "step": 24310
    },
    {
      "epoch": 12.7437106918239,
      "grad_norm": 0.4398140609264374,
      "learning_rate": 1.751071350665085e-05,
      "loss": 0.3926,
      "num_input_tokens_seen": 15902816,
      "step": 24315
    },
    {
      "epoch": 12.746331236897275,
      "grad_norm": 0.5150896310806274,
      "learning_rate": 1.7499805137945602e-05,
      "loss": 0.4656,
      "num_input_tokens_seen": 15905568,
      "step": 24320
    },
    {
      "epoch": 12.74895178197065,
      "grad_norm": 0.46541649103164673,
      "learning_rate": 1.7488898338195566e-05,
      "loss": 0.4121,
      "num_input_tokens_seen": 15909120,
      "step": 24325
    },
    {
      "epoch": 12.751572327044025,
      "grad_norm": 0.5241751670837402,
      "learning_rate": 1.7477993109682288e-05,
      "loss": 0.4175,
      "num_input_tokens_seen": 15911872,
      "step": 24330
    },
    {
      "epoch": 12.7541928721174,
      "grad_norm": 0.5024076700210571,
      "learning_rate": 1.7467089454687043e-05,
      "loss": 0.439,
      "num_input_tokens_seen": 15914784,
      "step": 24335
    },
    {
      "epoch": 12.756813417190775,
      "grad_norm": 0.34161707758903503,
      "learning_rate": 1.745618737549074e-05,
      "loss": 0.4419,
      "num_input_tokens_seen": 15918944,
      "step": 24340
    },
    {
      "epoch": 12.75943396226415,
      "grad_norm": 0.5659897327423096,
      "learning_rate": 1.7445286874373974e-05,
      "loss": 0.5116,
      "num_input_tokens_seen": 15921312,
      "step": 24345
    },
    {
      "epoch": 12.762054507337526,
      "grad_norm": 0.5427008867263794,
      "learning_rate": 1.743438795361701e-05,
      "loss": 0.4364,
      "num_input_tokens_seen": 15924320,
      "step": 24350
    },
    {
      "epoch": 12.764675052410901,
      "grad_norm": 0.4270116686820984,
      "learning_rate": 1.742349061549978e-05,
      "loss": 0.3909,
      "num_input_tokens_seen": 15927936,
      "step": 24355
    },
    {
      "epoch": 12.767295597484276,
      "grad_norm": 0.6756476759910583,
      "learning_rate": 1.741259486230186e-05,
      "loss": 0.6709,
      "num_input_tokens_seen": 15931104,
      "step": 24360
    },
    {
      "epoch": 12.769916142557651,
      "grad_norm": 1.2266401052474976,
      "learning_rate": 1.7401700696302546e-05,
      "loss": 0.499,
      "num_input_tokens_seen": 15935008,
      "step": 24365
    },
    {
      "epoch": 12.772536687631026,
      "grad_norm": 0.6647000312805176,
      "learning_rate": 1.739080811978075e-05,
      "loss": 0.4502,
      "num_input_tokens_seen": 15937824,
      "step": 24370
    },
    {
      "epoch": 12.775157232704402,
      "grad_norm": 0.40722861886024475,
      "learning_rate": 1.7379917135015093e-05,
      "loss": 0.3879,
      "num_input_tokens_seen": 15941056,
      "step": 24375
    },
    {
      "epoch": 12.777777777777779,
      "grad_norm": 0.48852837085723877,
      "learning_rate": 1.736902774428384e-05,
      "loss": 0.472,
      "num_input_tokens_seen": 15944640,
      "step": 24380
    },
    {
      "epoch": 12.780398322851154,
      "grad_norm": 0.509064793586731,
      "learning_rate": 1.735813994986491e-05,
      "loss": 0.3568,
      "num_input_tokens_seen": 15948384,
      "step": 24385
    },
    {
      "epoch": 12.783018867924529,
      "grad_norm": 0.9471514225006104,
      "learning_rate": 1.7347253754035933e-05,
      "loss": 0.3762,
      "num_input_tokens_seen": 15951232,
      "step": 24390
    },
    {
      "epoch": 12.785639412997904,
      "grad_norm": 0.5588351488113403,
      "learning_rate": 1.7336369159074154e-05,
      "loss": 0.4187,
      "num_input_tokens_seen": 15954816,
      "step": 24395
    },
    {
      "epoch": 12.78825995807128,
      "grad_norm": 0.36788323521614075,
      "learning_rate": 1.7325486167256522e-05,
      "loss": 0.6085,
      "num_input_tokens_seen": 15958336,
      "step": 24400
    },
    {
      "epoch": 12.790880503144654,
      "grad_norm": 0.2590753734111786,
      "learning_rate": 1.7314604780859623e-05,
      "loss": 0.4655,
      "num_input_tokens_seen": 15961984,
      "step": 24405
    },
    {
      "epoch": 12.79350104821803,
      "grad_norm": 0.5412381291389465,
      "learning_rate": 1.730372500215973e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 15965056,
      "step": 24410
    },
    {
      "epoch": 12.796121593291405,
      "grad_norm": 0.3950687348842621,
      "learning_rate": 1.7292846833432763e-05,
      "loss": 0.4488,
      "num_input_tokens_seen": 15967904,
      "step": 24415
    },
    {
      "epoch": 12.79874213836478,
      "grad_norm": 0.27247992157936096,
      "learning_rate": 1.7281970276954305e-05,
      "loss": 0.3915,
      "num_input_tokens_seen": 15971424,
      "step": 24420
    },
    {
      "epoch": 12.801362683438155,
      "grad_norm": 0.7592809796333313,
      "learning_rate": 1.727109533499961e-05,
      "loss": 0.6965,
      "num_input_tokens_seen": 15974272,
      "step": 24425
    },
    {
      "epoch": 12.80398322851153,
      "grad_norm": 0.5090323090553284,
      "learning_rate": 1.72602220098436e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 15977600,
      "step": 24430
    },
    {
      "epoch": 12.806603773584905,
      "grad_norm": 0.5596349239349365,
      "learning_rate": 1.7249350303760846e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 15980640,
      "step": 24435
    },
    {
      "epoch": 12.80922431865828,
      "grad_norm": 0.4030390977859497,
      "learning_rate": 1.7238480219025593e-05,
      "loss": 0.3949,
      "num_input_tokens_seen": 15983520,
      "step": 24440
    },
    {
      "epoch": 12.811844863731656,
      "grad_norm": 0.6433932781219482,
      "learning_rate": 1.7227611757911723e-05,
      "loss": 0.4874,
      "num_input_tokens_seen": 15986592,
      "step": 24445
    },
    {
      "epoch": 12.814465408805031,
      "grad_norm": 0.4001029431819916,
      "learning_rate": 1.7216744922692794e-05,
      "loss": 0.4829,
      "num_input_tokens_seen": 15989600,
      "step": 24450
    },
    {
      "epoch": 12.817085953878406,
      "grad_norm": 0.485027551651001,
      "learning_rate": 1.7205879715642037e-05,
      "loss": 0.4171,
      "num_input_tokens_seen": 15992416,
      "step": 24455
    },
    {
      "epoch": 12.819706498951781,
      "grad_norm": 0.6084916591644287,
      "learning_rate": 1.719501613903232e-05,
      "loss": 0.3508,
      "num_input_tokens_seen": 15994816,
      "step": 24460
    },
    {
      "epoch": 12.822327044025156,
      "grad_norm": 0.6082357168197632,
      "learning_rate": 1.7184154195136192e-05,
      "loss": 0.458,
      "num_input_tokens_seen": 15997536,
      "step": 24465
    },
    {
      "epoch": 12.824947589098532,
      "grad_norm": 0.3369195759296417,
      "learning_rate": 1.7173293886225837e-05,
      "loss": 0.417,
      "num_input_tokens_seen": 16001344,
      "step": 24470
    },
    {
      "epoch": 12.827568134171909,
      "grad_norm": 0.5790581703186035,
      "learning_rate": 1.71624352145731e-05,
      "loss": 0.3316,
      "num_input_tokens_seen": 16005440,
      "step": 24475
    },
    {
      "epoch": 12.830188679245284,
      "grad_norm": 0.39355605840682983,
      "learning_rate": 1.7151578182449508e-05,
      "loss": 0.4048,
      "num_input_tokens_seen": 16008992,
      "step": 24480
    },
    {
      "epoch": 12.832809224318659,
      "grad_norm": 0.40739476680755615,
      "learning_rate": 1.7140722792126212e-05,
      "loss": 0.4114,
      "num_input_tokens_seen": 16011712,
      "step": 24485
    },
    {
      "epoch": 12.835429769392034,
      "grad_norm": 0.3217387795448303,
      "learning_rate": 1.7129869045874053e-05,
      "loss": 0.4017,
      "num_input_tokens_seen": 16014528,
      "step": 24490
    },
    {
      "epoch": 12.83805031446541,
      "grad_norm": 0.42146340012550354,
      "learning_rate": 1.7119016945963506e-05,
      "loss": 0.3437,
      "num_input_tokens_seen": 16016736,
      "step": 24495
    },
    {
      "epoch": 12.840670859538784,
      "grad_norm": 0.7156957983970642,
      "learning_rate": 1.7108166494664694e-05,
      "loss": 0.3382,
      "num_input_tokens_seen": 16020128,
      "step": 24500
    },
    {
      "epoch": 12.84329140461216,
      "grad_norm": 0.40064892172813416,
      "learning_rate": 1.7097317694247417e-05,
      "loss": 0.3903,
      "num_input_tokens_seen": 16023008,
      "step": 24505
    },
    {
      "epoch": 12.845911949685535,
      "grad_norm": 0.33843010663986206,
      "learning_rate": 1.7086470546981122e-05,
      "loss": 0.3784,
      "num_input_tokens_seen": 16026656,
      "step": 24510
    },
    {
      "epoch": 12.84853249475891,
      "grad_norm": 0.698289692401886,
      "learning_rate": 1.7075625055134904e-05,
      "loss": 0.3976,
      "num_input_tokens_seen": 16029472,
      "step": 24515
    },
    {
      "epoch": 12.851153039832285,
      "grad_norm": 0.36824849247932434,
      "learning_rate": 1.706478122097752e-05,
      "loss": 0.4281,
      "num_input_tokens_seen": 16033504,
      "step": 24520
    },
    {
      "epoch": 12.85377358490566,
      "grad_norm": 0.5766488313674927,
      "learning_rate": 1.7053939046777384e-05,
      "loss": 0.3858,
      "num_input_tokens_seen": 16036480,
      "step": 24525
    },
    {
      "epoch": 12.856394129979035,
      "grad_norm": 0.6542719006538391,
      "learning_rate": 1.7043098534802533e-05,
      "loss": 0.5111,
      "num_input_tokens_seen": 16039072,
      "step": 24530
    },
    {
      "epoch": 12.85901467505241,
      "grad_norm": 0.4284506142139435,
      "learning_rate": 1.7032259687320695e-05,
      "loss": 0.4084,
      "num_input_tokens_seen": 16042144,
      "step": 24535
    },
    {
      "epoch": 12.861635220125786,
      "grad_norm": 0.4235589802265167,
      "learning_rate": 1.7021422506599228e-05,
      "loss": 0.4618,
      "num_input_tokens_seen": 16044672,
      "step": 24540
    },
    {
      "epoch": 12.864255765199161,
      "grad_norm": 0.5341126322746277,
      "learning_rate": 1.7010586994905155e-05,
      "loss": 0.4342,
      "num_input_tokens_seen": 16046912,
      "step": 24545
    },
    {
      "epoch": 12.866876310272536,
      "grad_norm": 0.42223799228668213,
      "learning_rate": 1.6999753154505123e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 16049920,
      "step": 24550
    },
    {
      "epoch": 12.869496855345911,
      "grad_norm": 0.3203638195991516,
      "learning_rate": 1.6988920987665476e-05,
      "loss": 0.3734,
      "num_input_tokens_seen": 16053952,
      "step": 24555
    },
    {
      "epoch": 12.872117400419286,
      "grad_norm": 0.6256056427955627,
      "learning_rate": 1.6978090496652156e-05,
      "loss": 0.4322,
      "num_input_tokens_seen": 16056352,
      "step": 24560
    },
    {
      "epoch": 12.874737945492662,
      "grad_norm": 0.4712960124015808,
      "learning_rate": 1.6967261683730778e-05,
      "loss": 0.4377,
      "num_input_tokens_seen": 16060608,
      "step": 24565
    },
    {
      "epoch": 12.877358490566039,
      "grad_norm": 0.25579455494880676,
      "learning_rate": 1.695643455116662e-05,
      "loss": 0.4592,
      "num_input_tokens_seen": 16063904,
      "step": 24570
    },
    {
      "epoch": 12.879979035639414,
      "grad_norm": 0.46955400705337524,
      "learning_rate": 1.694560910122458e-05,
      "loss": 0.3411,
      "num_input_tokens_seen": 16067136,
      "step": 24575
    },
    {
      "epoch": 12.882599580712789,
      "grad_norm": 0.7627587914466858,
      "learning_rate": 1.693478533616924e-05,
      "loss": 0.406,
      "num_input_tokens_seen": 16069696,
      "step": 24580
    },
    {
      "epoch": 12.885220125786164,
      "grad_norm": 0.45629769563674927,
      "learning_rate": 1.692396325826479e-05,
      "loss": 0.4535,
      "num_input_tokens_seen": 16072800,
      "step": 24585
    },
    {
      "epoch": 12.88784067085954,
      "grad_norm": 0.5919399857521057,
      "learning_rate": 1.6913142869775088e-05,
      "loss": 0.4798,
      "num_input_tokens_seen": 16075424,
      "step": 24590
    },
    {
      "epoch": 12.890461215932914,
      "grad_norm": 0.6086308360099792,
      "learning_rate": 1.6902324172963635e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 16078400,
      "step": 24595
    },
    {
      "epoch": 12.89308176100629,
      "grad_norm": 0.5903449058532715,
      "learning_rate": 1.6891507170093586e-05,
      "loss": 0.4207,
      "num_input_tokens_seen": 16086432,
      "step": 24600
    },
    {
      "epoch": 12.895702306079665,
      "grad_norm": 0.4551396071910858,
      "learning_rate": 1.6880691863427725e-05,
      "loss": 0.5124,
      "num_input_tokens_seen": 16089472,
      "step": 24605
    },
    {
      "epoch": 12.89832285115304,
      "grad_norm": 0.4586377441883087,
      "learning_rate": 1.686987825522851e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 16092800,
      "step": 24610
    },
    {
      "epoch": 12.900943396226415,
      "grad_norm": 0.37211135029792786,
      "learning_rate": 1.6859066347758002e-05,
      "loss": 0.3316,
      "num_input_tokens_seen": 16100992,
      "step": 24615
    },
    {
      "epoch": 12.90356394129979,
      "grad_norm": 0.5408166646957397,
      "learning_rate": 1.6848256143277933e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 16104320,
      "step": 24620
    },
    {
      "epoch": 12.906184486373165,
      "grad_norm": 0.4755297601222992,
      "learning_rate": 1.683744764404968e-05,
      "loss": 0.4913,
      "num_input_tokens_seen": 16108064,
      "step": 24625
    },
    {
      "epoch": 12.90880503144654,
      "grad_norm": 0.5001907348632812,
      "learning_rate": 1.6826640852334257e-05,
      "loss": 0.4654,
      "num_input_tokens_seen": 16111008,
      "step": 24630
    },
    {
      "epoch": 12.911425576519916,
      "grad_norm": 0.34808483719825745,
      "learning_rate": 1.6815835770392324e-05,
      "loss": 0.4949,
      "num_input_tokens_seen": 16114080,
      "step": 24635
    },
    {
      "epoch": 12.914046121593291,
      "grad_norm": 0.48104846477508545,
      "learning_rate": 1.6805032400484184e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 16116832,
      "step": 24640
    },
    {
      "epoch": 12.916666666666666,
      "grad_norm": 0.4815753698348999,
      "learning_rate": 1.6794230744869767e-05,
      "loss": 0.3661,
      "num_input_tokens_seen": 16119872,
      "step": 24645
    },
    {
      "epoch": 12.919287211740041,
      "grad_norm": 0.5105080604553223,
      "learning_rate": 1.6783430805808658e-05,
      "loss": 0.5311,
      "num_input_tokens_seen": 16123200,
      "step": 24650
    },
    {
      "epoch": 12.921907756813416,
      "grad_norm": 0.7056457996368408,
      "learning_rate": 1.6772632585560084e-05,
      "loss": 0.4157,
      "num_input_tokens_seen": 16126048,
      "step": 24655
    },
    {
      "epoch": 12.924528301886792,
      "grad_norm": 0.42834150791168213,
      "learning_rate": 1.6761836086382908e-05,
      "loss": 0.5198,
      "num_input_tokens_seen": 16129088,
      "step": 24660
    },
    {
      "epoch": 12.927148846960169,
      "grad_norm": 0.5829991698265076,
      "learning_rate": 1.6751041310535653e-05,
      "loss": 0.56,
      "num_input_tokens_seen": 16131584,
      "step": 24665
    },
    {
      "epoch": 12.929769392033544,
      "grad_norm": 0.6489884853363037,
      "learning_rate": 1.6740248260276437e-05,
      "loss": 0.4795,
      "num_input_tokens_seen": 16133984,
      "step": 24670
    },
    {
      "epoch": 12.932389937106919,
      "grad_norm": 0.5054169297218323,
      "learning_rate": 1.6729456937863047e-05,
      "loss": 0.4462,
      "num_input_tokens_seen": 16136480,
      "step": 24675
    },
    {
      "epoch": 12.935010482180294,
      "grad_norm": 0.5208187699317932,
      "learning_rate": 1.6718667345552918e-05,
      "loss": 0.3554,
      "num_input_tokens_seen": 16139616,
      "step": 24680
    },
    {
      "epoch": 12.93763102725367,
      "grad_norm": 0.4464419186115265,
      "learning_rate": 1.6707879485603097e-05,
      "loss": 0.5261,
      "num_input_tokens_seen": 16143008,
      "step": 24685
    },
    {
      "epoch": 12.940251572327044,
      "grad_norm": 0.5804262757301331,
      "learning_rate": 1.669709336027029e-05,
      "loss": 0.5047,
      "num_input_tokens_seen": 16146304,
      "step": 24690
    },
    {
      "epoch": 12.94287211740042,
      "grad_norm": 1.0110769271850586,
      "learning_rate": 1.6686308971810833e-05,
      "loss": 0.4537,
      "num_input_tokens_seen": 16149184,
      "step": 24695
    },
    {
      "epoch": 12.945492662473795,
      "grad_norm": 0.41079047322273254,
      "learning_rate": 1.6675526322480677e-05,
      "loss": 0.5337,
      "num_input_tokens_seen": 16151968,
      "step": 24700
    },
    {
      "epoch": 12.94811320754717,
      "grad_norm": 0.7135437726974487,
      "learning_rate": 1.6664745414535454e-05,
      "loss": 0.4616,
      "num_input_tokens_seen": 16154848,
      "step": 24705
    },
    {
      "epoch": 12.950733752620545,
      "grad_norm": 0.33617541193962097,
      "learning_rate": 1.665396625023039e-05,
      "loss": 0.4285,
      "num_input_tokens_seen": 16158336,
      "step": 24710
    },
    {
      "epoch": 12.95335429769392,
      "grad_norm": 0.9071377515792847,
      "learning_rate": 1.6643188831820375e-05,
      "loss": 0.5553,
      "num_input_tokens_seen": 16161856,
      "step": 24715
    },
    {
      "epoch": 12.955974842767295,
      "grad_norm": 0.47121474146842957,
      "learning_rate": 1.6632413161559912e-05,
      "loss": 0.4292,
      "num_input_tokens_seen": 16165696,
      "step": 24720
    },
    {
      "epoch": 12.95859538784067,
      "grad_norm": 0.8422400951385498,
      "learning_rate": 1.6621639241703162e-05,
      "loss": 0.4108,
      "num_input_tokens_seen": 16169440,
      "step": 24725
    },
    {
      "epoch": 12.961215932914046,
      "grad_norm": 0.293940007686615,
      "learning_rate": 1.6610867074503893e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 16173088,
      "step": 24730
    },
    {
      "epoch": 12.963836477987421,
      "grad_norm": 0.4855440855026245,
      "learning_rate": 1.660009666221552e-05,
      "loss": 0.4544,
      "num_input_tokens_seen": 16176352,
      "step": 24735
    },
    {
      "epoch": 12.966457023060796,
      "grad_norm": 0.6155287027359009,
      "learning_rate": 1.6589328007091093e-05,
      "loss": 0.3036,
      "num_input_tokens_seen": 16180128,
      "step": 24740
    },
    {
      "epoch": 12.969077568134171,
      "grad_norm": 0.48994025588035583,
      "learning_rate": 1.65785611113833e-05,
      "loss": 0.5401,
      "num_input_tokens_seen": 16182880,
      "step": 24745
    },
    {
      "epoch": 12.971698113207546,
      "grad_norm": 0.32185065746307373,
      "learning_rate": 1.6567795977344447e-05,
      "loss": 0.4214,
      "num_input_tokens_seen": 16186880,
      "step": 24750
    },
    {
      "epoch": 12.974318658280922,
      "grad_norm": 0.5816504955291748,
      "learning_rate": 1.6557032607226466e-05,
      "loss": 0.4773,
      "num_input_tokens_seen": 16189600,
      "step": 24755
    },
    {
      "epoch": 12.976939203354299,
      "grad_norm": 0.6832979917526245,
      "learning_rate": 1.654627100328095e-05,
      "loss": 0.5997,
      "num_input_tokens_seen": 16194976,
      "step": 24760
    },
    {
      "epoch": 12.979559748427674,
      "grad_norm": 0.45568975806236267,
      "learning_rate": 1.6535511167759085e-05,
      "loss": 0.5355,
      "num_input_tokens_seen": 16197536,
      "step": 24765
    },
    {
      "epoch": 12.982180293501049,
      "grad_norm": 0.5065709352493286,
      "learning_rate": 1.652475310291172e-05,
      "loss": 0.4995,
      "num_input_tokens_seen": 16201440,
      "step": 24770
    },
    {
      "epoch": 12.984800838574424,
      "grad_norm": 0.5529087781906128,
      "learning_rate": 1.6513996810989317e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 16204448,
      "step": 24775
    },
    {
      "epoch": 12.9874213836478,
      "grad_norm": 0.3750316798686981,
      "learning_rate": 1.6503242294241973e-05,
      "loss": 0.3541,
      "num_input_tokens_seen": 16207584,
      "step": 24780
    },
    {
      "epoch": 12.990041928721174,
      "grad_norm": 0.4661678671836853,
      "learning_rate": 1.6492489554919404e-05,
      "loss": 0.5139,
      "num_input_tokens_seen": 16209920,
      "step": 24785
    },
    {
      "epoch": 12.99266247379455,
      "grad_norm": 0.45783859491348267,
      "learning_rate": 1.6481738595270953e-05,
      "loss": 0.3988,
      "num_input_tokens_seen": 16212736,
      "step": 24790
    },
    {
      "epoch": 12.995283018867925,
      "grad_norm": 0.45548367500305176,
      "learning_rate": 1.6470989417545608e-05,
      "loss": 0.5241,
      "num_input_tokens_seen": 16216480,
      "step": 24795
    },
    {
      "epoch": 12.9979035639413,
      "grad_norm": 0.982658326625824,
      "learning_rate": 1.646024202399197e-05,
      "loss": 0.5622,
      "num_input_tokens_seen": 16218976,
      "step": 24800
    },
    {
      "epoch": 13.0,
      "eval_loss": 0.4753713011741638,
      "eval_runtime": 9.8451,
      "eval_samples_per_second": 86.134,
      "eval_steps_per_second": 21.534,
      "num_input_tokens_seen": 16220808,
      "step": 24804
    },
    {
      "epoch": 13.000524109014675,
      "grad_norm": 0.5339261293411255,
      "learning_rate": 1.6449496416858284e-05,
      "loss": 0.4194,
      "num_input_tokens_seen": 16221160,
      "step": 24805
    },
    {
      "epoch": 13.00314465408805,
      "grad_norm": 0.48718729615211487,
      "learning_rate": 1.64387525983924e-05,
      "loss": 0.5519,
      "num_input_tokens_seen": 16223848,
      "step": 24810
    },
    {
      "epoch": 13.005765199161425,
      "grad_norm": 0.23820769786834717,
      "learning_rate": 1.6428010570841788e-05,
      "loss": 0.3855,
      "num_input_tokens_seen": 16226856,
      "step": 24815
    },
    {
      "epoch": 13.0083857442348,
      "grad_norm": 0.42810317873954773,
      "learning_rate": 1.6417270336453573e-05,
      "loss": 0.3659,
      "num_input_tokens_seen": 16229736,
      "step": 24820
    },
    {
      "epoch": 13.011006289308176,
      "grad_norm": 0.7060415744781494,
      "learning_rate": 1.6406531897474492e-05,
      "loss": 0.3653,
      "num_input_tokens_seen": 16233064,
      "step": 24825
    },
    {
      "epoch": 13.01362683438155,
      "grad_norm": 0.4014207124710083,
      "learning_rate": 1.639579525615089e-05,
      "loss": 0.3888,
      "num_input_tokens_seen": 16236456,
      "step": 24830
    },
    {
      "epoch": 13.016247379454926,
      "grad_norm": 0.6431887149810791,
      "learning_rate": 1.638506041472877e-05,
      "loss": 0.6756,
      "num_input_tokens_seen": 16239272,
      "step": 24835
    },
    {
      "epoch": 13.018867924528301,
      "grad_norm": 0.7806167602539062,
      "learning_rate": 1.6374327375453724e-05,
      "loss": 0.3189,
      "num_input_tokens_seen": 16242152,
      "step": 24840
    },
    {
      "epoch": 13.021488469601676,
      "grad_norm": 0.5054982304573059,
      "learning_rate": 1.6363596140570982e-05,
      "loss": 0.46,
      "num_input_tokens_seen": 16245832,
      "step": 24845
    },
    {
      "epoch": 13.024109014675052,
      "grad_norm": 0.3039952516555786,
      "learning_rate": 1.6352866712325395e-05,
      "loss": 0.3551,
      "num_input_tokens_seen": 16249064,
      "step": 24850
    },
    {
      "epoch": 13.026729559748428,
      "grad_norm": 0.5749531388282776,
      "learning_rate": 1.634213909296144e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 16251944,
      "step": 24855
    },
    {
      "epoch": 13.029350104821804,
      "grad_norm": 0.4057743549346924,
      "learning_rate": 1.6331413284723217e-05,
      "loss": 0.5058,
      "num_input_tokens_seen": 16255368,
      "step": 24860
    },
    {
      "epoch": 13.031970649895179,
      "grad_norm": 0.6246649622917175,
      "learning_rate": 1.6320689289854442e-05,
      "loss": 0.476,
      "num_input_tokens_seen": 16258152,
      "step": 24865
    },
    {
      "epoch": 13.034591194968554,
      "grad_norm": 0.268045574426651,
      "learning_rate": 1.6309967110598438e-05,
      "loss": 0.452,
      "num_input_tokens_seen": 16261672,
      "step": 24870
    },
    {
      "epoch": 13.03721174004193,
      "grad_norm": 0.6709665060043335,
      "learning_rate": 1.6299246749198183e-05,
      "loss": 0.4647,
      "num_input_tokens_seen": 16265416,
      "step": 24875
    },
    {
      "epoch": 13.039832285115304,
      "grad_norm": 0.40833914279937744,
      "learning_rate": 1.6288528207896236e-05,
      "loss": 0.4286,
      "num_input_tokens_seen": 16268008,
      "step": 24880
    },
    {
      "epoch": 13.04245283018868,
      "grad_norm": 0.3623216152191162,
      "learning_rate": 1.627781148893481e-05,
      "loss": 0.304,
      "num_input_tokens_seen": 16271912,
      "step": 24885
    },
    {
      "epoch": 13.045073375262055,
      "grad_norm": 0.26680245995521545,
      "learning_rate": 1.6267096594555703e-05,
      "loss": 0.4519,
      "num_input_tokens_seen": 16275304,
      "step": 24890
    },
    {
      "epoch": 13.04769392033543,
      "grad_norm": 0.2539505958557129,
      "learning_rate": 1.6256383527000372e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 16279240,
      "step": 24895
    },
    {
      "epoch": 13.050314465408805,
      "grad_norm": 0.4714773893356323,
      "learning_rate": 1.6245672288509843e-05,
      "loss": 0.6609,
      "num_input_tokens_seen": 16282376,
      "step": 24900
    },
    {
      "epoch": 13.05293501048218,
      "grad_norm": 0.43410250544548035,
      "learning_rate": 1.62349628813248e-05,
      "loss": 0.4036,
      "num_input_tokens_seen": 16285000,
      "step": 24905
    },
    {
      "epoch": 13.055555555555555,
      "grad_norm": 0.5534791350364685,
      "learning_rate": 1.6224255307685525e-05,
      "loss": 0.464,
      "num_input_tokens_seen": 16287976,
      "step": 24910
    },
    {
      "epoch": 13.05817610062893,
      "grad_norm": 0.2886936068534851,
      "learning_rate": 1.6213549569831922e-05,
      "loss": 0.3797,
      "num_input_tokens_seen": 16291080,
      "step": 24915
    },
    {
      "epoch": 13.060796645702306,
      "grad_norm": 0.633669376373291,
      "learning_rate": 1.62028456700035e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 16294504,
      "step": 24920
    },
    {
      "epoch": 13.06341719077568,
      "grad_norm": 0.32670336961746216,
      "learning_rate": 1.619214361043942e-05,
      "loss": 0.4059,
      "num_input_tokens_seen": 16298696,
      "step": 24925
    },
    {
      "epoch": 13.066037735849056,
      "grad_norm": 0.35518667101860046,
      "learning_rate": 1.61814433933784e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 16302536,
      "step": 24930
    },
    {
      "epoch": 13.068658280922431,
      "grad_norm": 0.20020487904548645,
      "learning_rate": 1.617074502105881e-05,
      "loss": 0.3824,
      "num_input_tokens_seen": 16305992,
      "step": 24935
    },
    {
      "epoch": 13.071278825995806,
      "grad_norm": 0.459482878446579,
      "learning_rate": 1.616004849571864e-05,
      "loss": 0.5287,
      "num_input_tokens_seen": 16309768,
      "step": 24940
    },
    {
      "epoch": 13.073899371069182,
      "grad_norm": 0.36273565888404846,
      "learning_rate": 1.6149353819595463e-05,
      "loss": 0.3902,
      "num_input_tokens_seen": 16313160,
      "step": 24945
    },
    {
      "epoch": 13.076519916142558,
      "grad_norm": 0.5566925406455994,
      "learning_rate": 1.6138660994926508e-05,
      "loss": 0.3549,
      "num_input_tokens_seen": 16315688,
      "step": 24950
    },
    {
      "epoch": 13.079140461215934,
      "grad_norm": 1.375710368156433,
      "learning_rate": 1.612797002394857e-05,
      "loss": 0.4659,
      "num_input_tokens_seen": 16319208,
      "step": 24955
    },
    {
      "epoch": 13.081761006289309,
      "grad_norm": 0.37465861439704895,
      "learning_rate": 1.611728090889808e-05,
      "loss": 0.4106,
      "num_input_tokens_seen": 16322600,
      "step": 24960
    },
    {
      "epoch": 13.084381551362684,
      "grad_norm": 0.4612462520599365,
      "learning_rate": 1.610659365201109e-05,
      "loss": 0.4201,
      "num_input_tokens_seen": 16326568,
      "step": 24965
    },
    {
      "epoch": 13.08700209643606,
      "grad_norm": 0.5764386653900146,
      "learning_rate": 1.609590825552324e-05,
      "loss": 0.6479,
      "num_input_tokens_seen": 16331336,
      "step": 24970
    },
    {
      "epoch": 13.089622641509434,
      "grad_norm": 0.7016175389289856,
      "learning_rate": 1.6085224721669807e-05,
      "loss": 0.4584,
      "num_input_tokens_seen": 16334824,
      "step": 24975
    },
    {
      "epoch": 13.09224318658281,
      "grad_norm": 0.769939661026001,
      "learning_rate": 1.607454305268566e-05,
      "loss": 0.3786,
      "num_input_tokens_seen": 16337992,
      "step": 24980
    },
    {
      "epoch": 13.094863731656185,
      "grad_norm": 0.6270837783813477,
      "learning_rate": 1.606386325080528e-05,
      "loss": 0.5643,
      "num_input_tokens_seen": 16340680,
      "step": 24985
    },
    {
      "epoch": 13.09748427672956,
      "grad_norm": 0.43799731135368347,
      "learning_rate": 1.6053185318262746e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 16343464,
      "step": 24990
    },
    {
      "epoch": 13.100104821802935,
      "grad_norm": 0.5124051570892334,
      "learning_rate": 1.6042509257291783e-05,
      "loss": 0.4589,
      "num_input_tokens_seen": 16346376,
      "step": 24995
    },
    {
      "epoch": 13.10272536687631,
      "grad_norm": 0.32409390807151794,
      "learning_rate": 1.6031835070125685e-05,
      "loss": 0.2984,
      "num_input_tokens_seen": 16349032,
      "step": 25000
    },
    {
      "epoch": 13.105345911949685,
      "grad_norm": 0.40318140387535095,
      "learning_rate": 1.6021162758997382e-05,
      "loss": 0.5153,
      "num_input_tokens_seen": 16352104,
      "step": 25005
    },
    {
      "epoch": 13.10796645702306,
      "grad_norm": 0.5390914678573608,
      "learning_rate": 1.6010492326139398e-05,
      "loss": 0.5109,
      "num_input_tokens_seen": 16355496,
      "step": 25010
    },
    {
      "epoch": 13.110587002096436,
      "grad_norm": 0.3773222267627716,
      "learning_rate": 1.599982377378385e-05,
      "loss": 0.5005,
      "num_input_tokens_seen": 16359080,
      "step": 25015
    },
    {
      "epoch": 13.11320754716981,
      "grad_norm": 0.5103345513343811,
      "learning_rate": 1.5989157104162496e-05,
      "loss": 0.4585,
      "num_input_tokens_seen": 16362408,
      "step": 25020
    },
    {
      "epoch": 13.115828092243186,
      "grad_norm": 0.6419647336006165,
      "learning_rate": 1.5978492319506673e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 16365384,
      "step": 25025
    },
    {
      "epoch": 13.118448637316561,
      "grad_norm": 0.3084776997566223,
      "learning_rate": 1.596782942204733e-05,
      "loss": 0.3341,
      "num_input_tokens_seen": 16368840,
      "step": 25030
    },
    {
      "epoch": 13.121069182389936,
      "grad_norm": 0.5827191472053528,
      "learning_rate": 1.595716841401504e-05,
      "loss": 0.4119,
      "num_input_tokens_seen": 16372296,
      "step": 25035
    },
    {
      "epoch": 13.123689727463312,
      "grad_norm": 0.7144164443016052,
      "learning_rate": 1.594650929763994e-05,
      "loss": 0.31,
      "num_input_tokens_seen": 16375400,
      "step": 25040
    },
    {
      "epoch": 13.126310272536688,
      "grad_norm": 0.34149298071861267,
      "learning_rate": 1.5935852075151812e-05,
      "loss": 0.4801,
      "num_input_tokens_seen": 16378728,
      "step": 25045
    },
    {
      "epoch": 13.128930817610064,
      "grad_norm": 0.4925053119659424,
      "learning_rate": 1.5925196748780016e-05,
      "loss": 0.451,
      "num_input_tokens_seen": 16381672,
      "step": 25050
    },
    {
      "epoch": 13.131551362683439,
      "grad_norm": 0.6316483020782471,
      "learning_rate": 1.5914543320753528e-05,
      "loss": 0.5483,
      "num_input_tokens_seen": 16384744,
      "step": 25055
    },
    {
      "epoch": 13.134171907756814,
      "grad_norm": 0.5289615988731384,
      "learning_rate": 1.5903891793300932e-05,
      "loss": 0.57,
      "num_input_tokens_seen": 16387816,
      "step": 25060
    },
    {
      "epoch": 13.13679245283019,
      "grad_norm": 0.47110143303871155,
      "learning_rate": 1.5893242168650403e-05,
      "loss": 0.4233,
      "num_input_tokens_seen": 16390600,
      "step": 25065
    },
    {
      "epoch": 13.139412997903564,
      "grad_norm": 0.39095476269721985,
      "learning_rate": 1.5882594449029705e-05,
      "loss": 0.4231,
      "num_input_tokens_seen": 16394280,
      "step": 25070
    },
    {
      "epoch": 13.14203354297694,
      "grad_norm": 0.5153303146362305,
      "learning_rate": 1.587194863666624e-05,
      "loss": 0.4351,
      "num_input_tokens_seen": 16396904,
      "step": 25075
    },
    {
      "epoch": 13.144654088050315,
      "grad_norm": 0.5600583553314209,
      "learning_rate": 1.5861304733786974e-05,
      "loss": 0.4894,
      "num_input_tokens_seen": 16400136,
      "step": 25080
    },
    {
      "epoch": 13.14727463312369,
      "grad_norm": 0.4107963442802429,
      "learning_rate": 1.5850662742618505e-05,
      "loss": 0.3792,
      "num_input_tokens_seen": 16402568,
      "step": 25085
    },
    {
      "epoch": 13.149895178197065,
      "grad_norm": 0.3491901159286499,
      "learning_rate": 1.5840022665387004e-05,
      "loss": 0.5247,
      "num_input_tokens_seen": 16405832,
      "step": 25090
    },
    {
      "epoch": 13.15251572327044,
      "grad_norm": 0.8449594974517822,
      "learning_rate": 1.5829384504318268e-05,
      "loss": 0.6998,
      "num_input_tokens_seen": 16409288,
      "step": 25095
    },
    {
      "epoch": 13.155136268343815,
      "grad_norm": 0.42295515537261963,
      "learning_rate": 1.5818748261637662e-05,
      "loss": 0.3777,
      "num_input_tokens_seen": 16412584,
      "step": 25100
    },
    {
      "epoch": 13.15775681341719,
      "grad_norm": 0.38072124123573303,
      "learning_rate": 1.5808113939570173e-05,
      "loss": 0.4066,
      "num_input_tokens_seen": 16415080,
      "step": 25105
    },
    {
      "epoch": 13.160377358490566,
      "grad_norm": 0.4761376678943634,
      "learning_rate": 1.5797481540340386e-05,
      "loss": 0.4034,
      "num_input_tokens_seen": 16418184,
      "step": 25110
    },
    {
      "epoch": 13.16299790356394,
      "grad_norm": 0.41816624999046326,
      "learning_rate": 1.5786851066172466e-05,
      "loss": 0.6113,
      "num_input_tokens_seen": 16421672,
      "step": 25115
    },
    {
      "epoch": 13.165618448637316,
      "grad_norm": 0.37264230847358704,
      "learning_rate": 1.5776222519290208e-05,
      "loss": 0.4031,
      "num_input_tokens_seen": 16425448,
      "step": 25120
    },
    {
      "epoch": 13.168238993710691,
      "grad_norm": 0.29757821559906006,
      "learning_rate": 1.576559590191696e-05,
      "loss": 0.5154,
      "num_input_tokens_seen": 16428872,
      "step": 25125
    },
    {
      "epoch": 13.170859538784066,
      "grad_norm": 0.5829366445541382,
      "learning_rate": 1.5754971216275692e-05,
      "loss": 0.61,
      "num_input_tokens_seen": 16431560,
      "step": 25130
    },
    {
      "epoch": 13.173480083857442,
      "grad_norm": 0.5973299145698547,
      "learning_rate": 1.5744348464588975e-05,
      "loss": 0.454,
      "num_input_tokens_seen": 16435048,
      "step": 25135
    },
    {
      "epoch": 13.176100628930818,
      "grad_norm": 0.4512464106082916,
      "learning_rate": 1.5733727649078968e-05,
      "loss": 0.3938,
      "num_input_tokens_seen": 16438248,
      "step": 25140
    },
    {
      "epoch": 13.178721174004194,
      "grad_norm": 0.45014509558677673,
      "learning_rate": 1.5723108771967414e-05,
      "loss": 0.4683,
      "num_input_tokens_seen": 16441384,
      "step": 25145
    },
    {
      "epoch": 13.181341719077569,
      "grad_norm": 0.4884025454521179,
      "learning_rate": 1.5712491835475685e-05,
      "loss": 0.577,
      "num_input_tokens_seen": 16445992,
      "step": 25150
    },
    {
      "epoch": 13.183962264150944,
      "grad_norm": 0.2404436320066452,
      "learning_rate": 1.5701876841824698e-05,
      "loss": 0.4095,
      "num_input_tokens_seen": 16449416,
      "step": 25155
    },
    {
      "epoch": 13.18658280922432,
      "grad_norm": 0.40931403636932373,
      "learning_rate": 1.5691263793234993e-05,
      "loss": 0.3525,
      "num_input_tokens_seen": 16454472,
      "step": 25160
    },
    {
      "epoch": 13.189203354297694,
      "grad_norm": 0.5289860367774963,
      "learning_rate": 1.5680652691926706e-05,
      "loss": 0.4433,
      "num_input_tokens_seen": 16456808,
      "step": 25165
    },
    {
      "epoch": 13.19182389937107,
      "grad_norm": 0.9315758943557739,
      "learning_rate": 1.5670043540119553e-05,
      "loss": 0.5162,
      "num_input_tokens_seen": 16459752,
      "step": 25170
    },
    {
      "epoch": 13.194444444444445,
      "grad_norm": 0.5199068188667297,
      "learning_rate": 1.5659436340032855e-05,
      "loss": 0.4521,
      "num_input_tokens_seen": 16462760,
      "step": 25175
    },
    {
      "epoch": 13.19706498951782,
      "grad_norm": 0.4702714681625366,
      "learning_rate": 1.5648831093885515e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 16466536,
      "step": 25180
    },
    {
      "epoch": 13.199685534591195,
      "grad_norm": 0.5055330395698547,
      "learning_rate": 1.5638227803896015e-05,
      "loss": 0.6467,
      "num_input_tokens_seen": 16469384,
      "step": 25185
    },
    {
      "epoch": 13.20230607966457,
      "grad_norm": 0.5924569964408875,
      "learning_rate": 1.562762647228246e-05,
      "loss": 0.4808,
      "num_input_tokens_seen": 16472488,
      "step": 25190
    },
    {
      "epoch": 13.204926624737945,
      "grad_norm": 0.4505685269832611,
      "learning_rate": 1.5617027101262515e-05,
      "loss": 0.4686,
      "num_input_tokens_seen": 16475528,
      "step": 25195
    },
    {
      "epoch": 13.20754716981132,
      "grad_norm": 0.38869941234588623,
      "learning_rate": 1.5606429693053463e-05,
      "loss": 0.3411,
      "num_input_tokens_seen": 16478216,
      "step": 25200
    },
    {
      "epoch": 13.210167714884696,
      "grad_norm": 0.42446404695510864,
      "learning_rate": 1.5595834249872145e-05,
      "loss": 0.5146,
      "num_input_tokens_seen": 16481480,
      "step": 25205
    },
    {
      "epoch": 13.21278825995807,
      "grad_norm": 0.4429677426815033,
      "learning_rate": 1.558524077393502e-05,
      "loss": 0.4107,
      "num_input_tokens_seen": 16485768,
      "step": 25210
    },
    {
      "epoch": 13.215408805031446,
      "grad_norm": 0.3676183521747589,
      "learning_rate": 1.557464926745811e-05,
      "loss": 0.3693,
      "num_input_tokens_seen": 16490152,
      "step": 25215
    },
    {
      "epoch": 13.218029350104821,
      "grad_norm": 1.1211659908294678,
      "learning_rate": 1.5564059732657045e-05,
      "loss": 0.4123,
      "num_input_tokens_seen": 16493096,
      "step": 25220
    },
    {
      "epoch": 13.220649895178196,
      "grad_norm": 0.8739950060844421,
      "learning_rate": 1.555347217174703e-05,
      "loss": 0.4111,
      "num_input_tokens_seen": 16495848,
      "step": 25225
    },
    {
      "epoch": 13.223270440251572,
      "grad_norm": 0.3959674835205078,
      "learning_rate": 1.554288658694287e-05,
      "loss": 0.4752,
      "num_input_tokens_seen": 16502888,
      "step": 25230
    },
    {
      "epoch": 13.225890985324948,
      "grad_norm": 0.6803134083747864,
      "learning_rate": 1.553230298045895e-05,
      "loss": 0.4399,
      "num_input_tokens_seen": 16505928,
      "step": 25235
    },
    {
      "epoch": 13.228511530398324,
      "grad_norm": 0.5247528553009033,
      "learning_rate": 1.5521721354509223e-05,
      "loss": 0.565,
      "num_input_tokens_seen": 16508648,
      "step": 25240
    },
    {
      "epoch": 13.231132075471699,
      "grad_norm": 0.4542468786239624,
      "learning_rate": 1.5511141711307258e-05,
      "loss": 0.4814,
      "num_input_tokens_seen": 16512968,
      "step": 25245
    },
    {
      "epoch": 13.233752620545074,
      "grad_norm": 0.2932594120502472,
      "learning_rate": 1.5500564053066184e-05,
      "loss": 0.5599,
      "num_input_tokens_seen": 16516552,
      "step": 25250
    },
    {
      "epoch": 13.23637316561845,
      "grad_norm": 0.6646981239318848,
      "learning_rate": 1.5489988381998743e-05,
      "loss": 0.3467,
      "num_input_tokens_seen": 16519528,
      "step": 25255
    },
    {
      "epoch": 13.238993710691824,
      "grad_norm": 0.48278534412384033,
      "learning_rate": 1.5479414700317237e-05,
      "loss": 0.4123,
      "num_input_tokens_seen": 16522504,
      "step": 25260
    },
    {
      "epoch": 13.2416142557652,
      "grad_norm": 0.748799741268158,
      "learning_rate": 1.546884301023356e-05,
      "loss": 0.5977,
      "num_input_tokens_seen": 16525384,
      "step": 25265
    },
    {
      "epoch": 13.244234800838575,
      "grad_norm": 0.6850560307502747,
      "learning_rate": 1.5458273313959186e-05,
      "loss": 0.5135,
      "num_input_tokens_seen": 16528840,
      "step": 25270
    },
    {
      "epoch": 13.24685534591195,
      "grad_norm": 0.6439611315727234,
      "learning_rate": 1.5447705613705177e-05,
      "loss": 0.4529,
      "num_input_tokens_seen": 16532648,
      "step": 25275
    },
    {
      "epoch": 13.249475890985325,
      "grad_norm": 0.4758899211883545,
      "learning_rate": 1.5437139911682176e-05,
      "loss": 0.4076,
      "num_input_tokens_seen": 16535720,
      "step": 25280
    },
    {
      "epoch": 13.2520964360587,
      "grad_norm": 0.6492986083030701,
      "learning_rate": 1.54265762101004e-05,
      "loss": 0.3893,
      "num_input_tokens_seen": 16538312,
      "step": 25285
    },
    {
      "epoch": 13.254716981132075,
      "grad_norm": 0.7447372078895569,
      "learning_rate": 1.5416014511169663e-05,
      "loss": 0.5645,
      "num_input_tokens_seen": 16541640,
      "step": 25290
    },
    {
      "epoch": 13.25733752620545,
      "grad_norm": 0.2776636779308319,
      "learning_rate": 1.5405454817099363e-05,
      "loss": 0.3842,
      "num_input_tokens_seen": 16545640,
      "step": 25295
    },
    {
      "epoch": 13.259958071278826,
      "grad_norm": 0.4620876610279083,
      "learning_rate": 1.5394897130098446e-05,
      "loss": 0.4454,
      "num_input_tokens_seen": 16548488,
      "step": 25300
    },
    {
      "epoch": 13.2625786163522,
      "grad_norm": 0.480864942073822,
      "learning_rate": 1.5384341452375465e-05,
      "loss": 0.4197,
      "num_input_tokens_seen": 16551592,
      "step": 25305
    },
    {
      "epoch": 13.265199161425576,
      "grad_norm": 0.48270586133003235,
      "learning_rate": 1.5373787786138556e-05,
      "loss": 0.4869,
      "num_input_tokens_seen": 16554600,
      "step": 25310
    },
    {
      "epoch": 13.267819706498951,
      "grad_norm": 0.522497832775116,
      "learning_rate": 1.5363236133595416e-05,
      "loss": 0.4826,
      "num_input_tokens_seen": 16557192,
      "step": 25315
    },
    {
      "epoch": 13.270440251572326,
      "grad_norm": 0.8091654181480408,
      "learning_rate": 1.5352686496953344e-05,
      "loss": 0.6077,
      "num_input_tokens_seen": 16560040,
      "step": 25320
    },
    {
      "epoch": 13.273060796645701,
      "grad_norm": 0.4108631908893585,
      "learning_rate": 1.534213887841919e-05,
      "loss": 0.399,
      "num_input_tokens_seen": 16563304,
      "step": 25325
    },
    {
      "epoch": 13.275681341719078,
      "grad_norm": 0.6888351440429688,
      "learning_rate": 1.5331593280199395e-05,
      "loss": 0.4227,
      "num_input_tokens_seen": 16566312,
      "step": 25330
    },
    {
      "epoch": 13.278301886792454,
      "grad_norm": 0.9886233806610107,
      "learning_rate": 1.532104970449999e-05,
      "loss": 0.4841,
      "num_input_tokens_seen": 16569736,
      "step": 25335
    },
    {
      "epoch": 13.280922431865829,
      "grad_norm": 0.46424850821495056,
      "learning_rate": 1.5310508153526555e-05,
      "loss": 0.4675,
      "num_input_tokens_seen": 16572680,
      "step": 25340
    },
    {
      "epoch": 13.283542976939204,
      "grad_norm": 0.4417768120765686,
      "learning_rate": 1.5299968629484276e-05,
      "loss": 0.4224,
      "num_input_tokens_seen": 16576168,
      "step": 25345
    },
    {
      "epoch": 13.286163522012579,
      "grad_norm": 0.9486600756645203,
      "learning_rate": 1.5289431134577904e-05,
      "loss": 0.4517,
      "num_input_tokens_seen": 16578504,
      "step": 25350
    },
    {
      "epoch": 13.288784067085954,
      "grad_norm": 0.4731731414794922,
      "learning_rate": 1.527889567101174e-05,
      "loss": 0.4893,
      "num_input_tokens_seen": 16582472,
      "step": 25355
    },
    {
      "epoch": 13.29140461215933,
      "grad_norm": 0.4409058094024658,
      "learning_rate": 1.5268362240989707e-05,
      "loss": 0.517,
      "num_input_tokens_seen": 16585128,
      "step": 25360
    },
    {
      "epoch": 13.294025157232705,
      "grad_norm": 0.48500871658325195,
      "learning_rate": 1.525783084671526e-05,
      "loss": 0.4958,
      "num_input_tokens_seen": 16587592,
      "step": 25365
    },
    {
      "epoch": 13.29664570230608,
      "grad_norm": 0.42459359765052795,
      "learning_rate": 1.5247301490391458e-05,
      "loss": 0.4468,
      "num_input_tokens_seen": 16594600,
      "step": 25370
    },
    {
      "epoch": 13.299266247379455,
      "grad_norm": 0.5011045932769775,
      "learning_rate": 1.5236774174220922e-05,
      "loss": 0.3307,
      "num_input_tokens_seen": 16598120,
      "step": 25375
    },
    {
      "epoch": 13.30188679245283,
      "grad_norm": 0.24990305304527283,
      "learning_rate": 1.5226248900405852e-05,
      "loss": 0.5533,
      "num_input_tokens_seen": 16601768,
      "step": 25380
    },
    {
      "epoch": 13.304507337526205,
      "grad_norm": 0.3408730924129486,
      "learning_rate": 1.5215725671147996e-05,
      "loss": 0.3535,
      "num_input_tokens_seen": 16604776,
      "step": 25385
    },
    {
      "epoch": 13.30712788259958,
      "grad_norm": 0.37993910908699036,
      "learning_rate": 1.5205204488648709e-05,
      "loss": 0.3883,
      "num_input_tokens_seen": 16607688,
      "step": 25390
    },
    {
      "epoch": 13.309748427672956,
      "grad_norm": 0.3372069001197815,
      "learning_rate": 1.5194685355108893e-05,
      "loss": 0.3795,
      "num_input_tokens_seen": 16610920,
      "step": 25395
    },
    {
      "epoch": 13.31236897274633,
      "grad_norm": 0.627095103263855,
      "learning_rate": 1.518416827272904e-05,
      "loss": 0.3498,
      "num_input_tokens_seen": 16614024,
      "step": 25400
    },
    {
      "epoch": 13.314989517819706,
      "grad_norm": 0.5020085573196411,
      "learning_rate": 1.5173653243709207e-05,
      "loss": 0.447,
      "num_input_tokens_seen": 16616936,
      "step": 25405
    },
    {
      "epoch": 13.317610062893081,
      "grad_norm": 0.4680784046649933,
      "learning_rate": 1.5163140270249e-05,
      "loss": 0.4265,
      "num_input_tokens_seen": 16619432,
      "step": 25410
    },
    {
      "epoch": 13.320230607966456,
      "grad_norm": 0.5692754983901978,
      "learning_rate": 1.5152629354547626e-05,
      "loss": 0.4,
      "num_input_tokens_seen": 16622152,
      "step": 25415
    },
    {
      "epoch": 13.322851153039831,
      "grad_norm": 0.7138769626617432,
      "learning_rate": 1.5142120498803842e-05,
      "loss": 0.4891,
      "num_input_tokens_seen": 16624808,
      "step": 25420
    },
    {
      "epoch": 13.325471698113208,
      "grad_norm": 0.3026657998561859,
      "learning_rate": 1.5131613705215991e-05,
      "loss": 0.5674,
      "num_input_tokens_seen": 16628744,
      "step": 25425
    },
    {
      "epoch": 13.328092243186584,
      "grad_norm": 0.46720027923583984,
      "learning_rate": 1.5121108975981965e-05,
      "loss": 0.5209,
      "num_input_tokens_seen": 16631368,
      "step": 25430
    },
    {
      "epoch": 13.330712788259959,
      "grad_norm": 0.424950510263443,
      "learning_rate": 1.5110606313299248e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 16634536,
      "step": 25435
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 0.46303972601890564,
      "learning_rate": 1.5100105719364855e-05,
      "loss": 0.4785,
      "num_input_tokens_seen": 16637544,
      "step": 25440
    },
    {
      "epoch": 13.335953878406709,
      "grad_norm": 0.4501211941242218,
      "learning_rate": 1.50896071963754e-05,
      "loss": 0.4278,
      "num_input_tokens_seen": 16641096,
      "step": 25445
    },
    {
      "epoch": 13.338574423480084,
      "grad_norm": 0.46587520837783813,
      "learning_rate": 1.5079110746527055e-05,
      "loss": 0.5148,
      "num_input_tokens_seen": 16643688,
      "step": 25450
    },
    {
      "epoch": 13.34119496855346,
      "grad_norm": 0.49915611743927,
      "learning_rate": 1.5068616372015565e-05,
      "loss": 0.4963,
      "num_input_tokens_seen": 16646888,
      "step": 25455
    },
    {
      "epoch": 13.343815513626835,
      "grad_norm": 0.3729545474052429,
      "learning_rate": 1.5058124075036223e-05,
      "loss": 0.3667,
      "num_input_tokens_seen": 16650312,
      "step": 25460
    },
    {
      "epoch": 13.34643605870021,
      "grad_norm": 0.47777312994003296,
      "learning_rate": 1.5047633857783908e-05,
      "loss": 0.4128,
      "num_input_tokens_seen": 16653896,
      "step": 25465
    },
    {
      "epoch": 13.349056603773585,
      "grad_norm": 0.6001657247543335,
      "learning_rate": 1.5037145722453044e-05,
      "loss": 0.3567,
      "num_input_tokens_seen": 16657320,
      "step": 25470
    },
    {
      "epoch": 13.35167714884696,
      "grad_norm": 0.3102976381778717,
      "learning_rate": 1.5026659671237624e-05,
      "loss": 0.4088,
      "num_input_tokens_seen": 16660456,
      "step": 25475
    },
    {
      "epoch": 13.354297693920335,
      "grad_norm": 0.3836674094200134,
      "learning_rate": 1.501617570633123e-05,
      "loss": 0.4511,
      "num_input_tokens_seen": 16664424,
      "step": 25480
    },
    {
      "epoch": 13.35691823899371,
      "grad_norm": 0.47960570454597473,
      "learning_rate": 1.500569382992697e-05,
      "loss": 0.5202,
      "num_input_tokens_seen": 16667656,
      "step": 25485
    },
    {
      "epoch": 13.359538784067086,
      "grad_norm": 0.4187663197517395,
      "learning_rate": 1.4995214044217547e-05,
      "loss": 0.4191,
      "num_input_tokens_seen": 16671688,
      "step": 25490
    },
    {
      "epoch": 13.36215932914046,
      "grad_norm": 0.3931616246700287,
      "learning_rate": 1.49847363513952e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 16674856,
      "step": 25495
    },
    {
      "epoch": 13.364779874213836,
      "grad_norm": 0.4703308343887329,
      "learning_rate": 1.4974260753651748e-05,
      "loss": 0.3831,
      "num_input_tokens_seen": 16677864,
      "step": 25500
    },
    {
      "epoch": 13.367400419287211,
      "grad_norm": 0.666793167591095,
      "learning_rate": 1.4963787253178568e-05,
      "loss": 0.3711,
      "num_input_tokens_seen": 16680808,
      "step": 25505
    },
    {
      "epoch": 13.370020964360586,
      "grad_norm": 1.3328280448913574,
      "learning_rate": 1.4953315852166594e-05,
      "loss": 0.3911,
      "num_input_tokens_seen": 16683208,
      "step": 25510
    },
    {
      "epoch": 13.372641509433961,
      "grad_norm": 0.360973060131073,
      "learning_rate": 1.4942846552806322e-05,
      "loss": 0.4223,
      "num_input_tokens_seen": 16686952,
      "step": 25515
    },
    {
      "epoch": 13.375262054507338,
      "grad_norm": 0.7665466070175171,
      "learning_rate": 1.4932379357287829e-05,
      "loss": 0.554,
      "num_input_tokens_seen": 16689576,
      "step": 25520
    },
    {
      "epoch": 13.377882599580714,
      "grad_norm": 0.32847920060157776,
      "learning_rate": 1.49219142678007e-05,
      "loss": 0.4276,
      "num_input_tokens_seen": 16693320,
      "step": 25525
    },
    {
      "epoch": 13.380503144654089,
      "grad_norm": 0.4908786714076996,
      "learning_rate": 1.4911451286534128e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 16698344,
      "step": 25530
    },
    {
      "epoch": 13.383123689727464,
      "grad_norm": 0.5270018577575684,
      "learning_rate": 1.490099041567686e-05,
      "loss": 0.4146,
      "num_input_tokens_seen": 16701576,
      "step": 25535
    },
    {
      "epoch": 13.385744234800839,
      "grad_norm": 0.9542834162712097,
      "learning_rate": 1.4890531657417176e-05,
      "loss": 0.4569,
      "num_input_tokens_seen": 16704904,
      "step": 25540
    },
    {
      "epoch": 13.388364779874214,
      "grad_norm": 0.37936317920684814,
      "learning_rate": 1.4880075013942938e-05,
      "loss": 0.4758,
      "num_input_tokens_seen": 16707304,
      "step": 25545
    },
    {
      "epoch": 13.39098532494759,
      "grad_norm": 0.4278789460659027,
      "learning_rate": 1.4869620487441558e-05,
      "loss": 0.5037,
      "num_input_tokens_seen": 16710120,
      "step": 25550
    },
    {
      "epoch": 13.393605870020965,
      "grad_norm": 0.3933885097503662,
      "learning_rate": 1.4859168080099988e-05,
      "loss": 0.3789,
      "num_input_tokens_seen": 16713736,
      "step": 25555
    },
    {
      "epoch": 13.39622641509434,
      "grad_norm": 0.5581837892532349,
      "learning_rate": 1.484871779410477e-05,
      "loss": 0.4171,
      "num_input_tokens_seen": 16716616,
      "step": 25560
    },
    {
      "epoch": 13.398846960167715,
      "grad_norm": 0.7026768326759338,
      "learning_rate": 1.4838269631641971e-05,
      "loss": 0.6261,
      "num_input_tokens_seen": 16719432,
      "step": 25565
    },
    {
      "epoch": 13.40146750524109,
      "grad_norm": 0.39154788851737976,
      "learning_rate": 1.482782359489724e-05,
      "loss": 0.4545,
      "num_input_tokens_seen": 16723176,
      "step": 25570
    },
    {
      "epoch": 13.404088050314465,
      "grad_norm": 0.8910542726516724,
      "learning_rate": 1.4817379686055754e-05,
      "loss": 0.5206,
      "num_input_tokens_seen": 16725608,
      "step": 25575
    },
    {
      "epoch": 13.40670859538784,
      "grad_norm": 0.6260226964950562,
      "learning_rate": 1.4806937907302282e-05,
      "loss": 0.4329,
      "num_input_tokens_seen": 16729320,
      "step": 25580
    },
    {
      "epoch": 13.409329140461216,
      "grad_norm": 0.491805762052536,
      "learning_rate": 1.4796498260821103e-05,
      "loss": 0.5101,
      "num_input_tokens_seen": 16731976,
      "step": 25585
    },
    {
      "epoch": 13.41194968553459,
      "grad_norm": 0.37352800369262695,
      "learning_rate": 1.4786060748796076e-05,
      "loss": 0.523,
      "num_input_tokens_seen": 16734792,
      "step": 25590
    },
    {
      "epoch": 13.414570230607966,
      "grad_norm": 0.3268832266330719,
      "learning_rate": 1.4775625373410617e-05,
      "loss": 0.5565,
      "num_input_tokens_seen": 16737608,
      "step": 25595
    },
    {
      "epoch": 13.417190775681341,
      "grad_norm": 0.5093991756439209,
      "learning_rate": 1.4765192136847685e-05,
      "loss": 0.3678,
      "num_input_tokens_seen": 16741000,
      "step": 25600
    },
    {
      "epoch": 13.419811320754716,
      "grad_norm": 0.6459836959838867,
      "learning_rate": 1.47547610412898e-05,
      "loss": 0.409,
      "num_input_tokens_seen": 16743944,
      "step": 25605
    },
    {
      "epoch": 13.422431865828091,
      "grad_norm": 0.4885968863964081,
      "learning_rate": 1.4744332088919011e-05,
      "loss": 0.4152,
      "num_input_tokens_seen": 16747176,
      "step": 25610
    },
    {
      "epoch": 13.425052410901468,
      "grad_norm": 0.43460530042648315,
      "learning_rate": 1.4733905281916954e-05,
      "loss": 0.4918,
      "num_input_tokens_seen": 16749928,
      "step": 25615
    },
    {
      "epoch": 13.427672955974844,
      "grad_norm": 0.38337379693984985,
      "learning_rate": 1.472348062246478e-05,
      "loss": 0.3775,
      "num_input_tokens_seen": 16753128,
      "step": 25620
    },
    {
      "epoch": 13.430293501048219,
      "grad_norm": 0.3852388858795166,
      "learning_rate": 1.4713058112743228e-05,
      "loss": 0.2804,
      "num_input_tokens_seen": 16756104,
      "step": 25625
    },
    {
      "epoch": 13.432914046121594,
      "grad_norm": 0.4310401976108551,
      "learning_rate": 1.4702637754932553e-05,
      "loss": 0.4767,
      "num_input_tokens_seen": 16758824,
      "step": 25630
    },
    {
      "epoch": 13.435534591194969,
      "grad_norm": 0.5117664933204651,
      "learning_rate": 1.4692219551212587e-05,
      "loss": 0.4453,
      "num_input_tokens_seen": 16761224,
      "step": 25635
    },
    {
      "epoch": 13.438155136268344,
      "grad_norm": 0.40952977538108826,
      "learning_rate": 1.4681803503762695e-05,
      "loss": 0.4658,
      "num_input_tokens_seen": 16764072,
      "step": 25640
    },
    {
      "epoch": 13.44077568134172,
      "grad_norm": 0.6498307585716248,
      "learning_rate": 1.4671389614761788e-05,
      "loss": 0.3574,
      "num_input_tokens_seen": 16767560,
      "step": 25645
    },
    {
      "epoch": 13.443396226415095,
      "grad_norm": 0.4346383213996887,
      "learning_rate": 1.4660977886388339e-05,
      "loss": 0.3628,
      "num_input_tokens_seen": 16770504,
      "step": 25650
    },
    {
      "epoch": 13.44601677148847,
      "grad_norm": 0.5649580359458923,
      "learning_rate": 1.4650568320820362e-05,
      "loss": 0.4625,
      "num_input_tokens_seen": 16773224,
      "step": 25655
    },
    {
      "epoch": 13.448637316561845,
      "grad_norm": 0.5312904715538025,
      "learning_rate": 1.4640160920235423e-05,
      "loss": 0.4413,
      "num_input_tokens_seen": 16775880,
      "step": 25660
    },
    {
      "epoch": 13.45125786163522,
      "grad_norm": 0.21722014248371124,
      "learning_rate": 1.4629755686810631e-05,
      "loss": 0.4648,
      "num_input_tokens_seen": 16778920,
      "step": 25665
    },
    {
      "epoch": 13.453878406708595,
      "grad_norm": 0.3706784248352051,
      "learning_rate": 1.4619352622722632e-05,
      "loss": 0.3826,
      "num_input_tokens_seen": 16781544,
      "step": 25670
    },
    {
      "epoch": 13.45649895178197,
      "grad_norm": 1.0158064365386963,
      "learning_rate": 1.4608951730147635e-05,
      "loss": 0.4749,
      "num_input_tokens_seen": 16784264,
      "step": 25675
    },
    {
      "epoch": 13.459119496855346,
      "grad_norm": 0.30534687638282776,
      "learning_rate": 1.4598553011261395e-05,
      "loss": 0.4327,
      "num_input_tokens_seen": 16787176,
      "step": 25680
    },
    {
      "epoch": 13.46174004192872,
      "grad_norm": 0.3697817027568817,
      "learning_rate": 1.4588156468239192e-05,
      "loss": 0.4274,
      "num_input_tokens_seen": 16790376,
      "step": 25685
    },
    {
      "epoch": 13.464360587002096,
      "grad_norm": 0.5656589865684509,
      "learning_rate": 1.4577762103255882e-05,
      "loss": 0.4095,
      "num_input_tokens_seen": 16793608,
      "step": 25690
    },
    {
      "epoch": 13.466981132075471,
      "grad_norm": 0.7041189074516296,
      "learning_rate": 1.4567369918485828e-05,
      "loss": 0.6363,
      "num_input_tokens_seen": 16796552,
      "step": 25695
    },
    {
      "epoch": 13.469601677148846,
      "grad_norm": 0.4663284420967102,
      "learning_rate": 1.4556979916102976e-05,
      "loss": 0.4954,
      "num_input_tokens_seen": 16799432,
      "step": 25700
    },
    {
      "epoch": 13.472222222222221,
      "grad_norm": 0.6730542778968811,
      "learning_rate": 1.454659209828077e-05,
      "loss": 0.4761,
      "num_input_tokens_seen": 16802184,
      "step": 25705
    },
    {
      "epoch": 13.474842767295598,
      "grad_norm": 1.4946917295455933,
      "learning_rate": 1.4536206467192243e-05,
      "loss": 0.2904,
      "num_input_tokens_seen": 16807208,
      "step": 25710
    },
    {
      "epoch": 13.477463312368974,
      "grad_norm": 0.6032067537307739,
      "learning_rate": 1.4525823025009944e-05,
      "loss": 0.4599,
      "num_input_tokens_seen": 16810856,
      "step": 25715
    },
    {
      "epoch": 13.480083857442349,
      "grad_norm": 0.5411683917045593,
      "learning_rate": 1.4515441773905983e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 16813736,
      "step": 25720
    },
    {
      "epoch": 13.482704402515724,
      "grad_norm": 0.4439781606197357,
      "learning_rate": 1.4505062716051988e-05,
      "loss": 0.3796,
      "num_input_tokens_seen": 16817928,
      "step": 25725
    },
    {
      "epoch": 13.485324947589099,
      "grad_norm": 0.42843759059906006,
      "learning_rate": 1.4494685853619133e-05,
      "loss": 0.3514,
      "num_input_tokens_seen": 16820872,
      "step": 25730
    },
    {
      "epoch": 13.487945492662474,
      "grad_norm": 0.5209152102470398,
      "learning_rate": 1.4484311188778143e-05,
      "loss": 0.4767,
      "num_input_tokens_seen": 16823464,
      "step": 25735
    },
    {
      "epoch": 13.49056603773585,
      "grad_norm": 0.3307298719882965,
      "learning_rate": 1.4473938723699285e-05,
      "loss": 0.4195,
      "num_input_tokens_seen": 16826632,
      "step": 25740
    },
    {
      "epoch": 13.493186582809225,
      "grad_norm": 1.056525707244873,
      "learning_rate": 1.4463568460552368e-05,
      "loss": 0.4519,
      "num_input_tokens_seen": 16829640,
      "step": 25745
    },
    {
      "epoch": 13.4958071278826,
      "grad_norm": 0.6146100759506226,
      "learning_rate": 1.4453200401506723e-05,
      "loss": 0.4124,
      "num_input_tokens_seen": 16832456,
      "step": 25750
    },
    {
      "epoch": 13.498427672955975,
      "grad_norm": 0.39370426535606384,
      "learning_rate": 1.4442834548731218e-05,
      "loss": 0.4504,
      "num_input_tokens_seen": 16835208,
      "step": 25755
    },
    {
      "epoch": 13.50104821802935,
      "grad_norm": 0.47574785351753235,
      "learning_rate": 1.4432470904394285e-05,
      "loss": 0.4478,
      "num_input_tokens_seen": 16838632,
      "step": 25760
    },
    {
      "epoch": 13.503668763102725,
      "grad_norm": 0.49706751108169556,
      "learning_rate": 1.4422109470663878e-05,
      "loss": 0.5005,
      "num_input_tokens_seen": 16842088,
      "step": 25765
    },
    {
      "epoch": 13.5062893081761,
      "grad_norm": 1.0798828601837158,
      "learning_rate": 1.441175024970749e-05,
      "loss": 0.3809,
      "num_input_tokens_seen": 16844488,
      "step": 25770
    },
    {
      "epoch": 13.508909853249476,
      "grad_norm": 1.5004048347473145,
      "learning_rate": 1.4401393243692163e-05,
      "loss": 0.5491,
      "num_input_tokens_seen": 16848712,
      "step": 25775
    },
    {
      "epoch": 13.51153039832285,
      "grad_norm": 0.49737244844436646,
      "learning_rate": 1.4391038454784455e-05,
      "loss": 0.371,
      "num_input_tokens_seen": 16852232,
      "step": 25780
    },
    {
      "epoch": 13.514150943396226,
      "grad_norm": 0.2702628970146179,
      "learning_rate": 1.4380685885150458e-05,
      "loss": 0.4259,
      "num_input_tokens_seen": 16855368,
      "step": 25785
    },
    {
      "epoch": 13.516771488469601,
      "grad_norm": 0.7356153726577759,
      "learning_rate": 1.437033553695582e-05,
      "loss": 0.4556,
      "num_input_tokens_seen": 16858536,
      "step": 25790
    },
    {
      "epoch": 13.519392033542976,
      "grad_norm": 0.4708496034145355,
      "learning_rate": 1.4359987412365721e-05,
      "loss": 0.409,
      "num_input_tokens_seen": 16862152,
      "step": 25795
    },
    {
      "epoch": 13.522012578616351,
      "grad_norm": 0.7962355017662048,
      "learning_rate": 1.4349641513544876e-05,
      "loss": 0.4318,
      "num_input_tokens_seen": 16864904,
      "step": 25800
    },
    {
      "epoch": 13.524633123689728,
      "grad_norm": 0.5495277643203735,
      "learning_rate": 1.4339297842657523e-05,
      "loss": 0.5172,
      "num_input_tokens_seen": 16868136,
      "step": 25805
    },
    {
      "epoch": 13.527253668763104,
      "grad_norm": 0.3930613100528717,
      "learning_rate": 1.4328956401867427e-05,
      "loss": 0.4496,
      "num_input_tokens_seen": 16871432,
      "step": 25810
    },
    {
      "epoch": 13.529874213836479,
      "grad_norm": 0.43685829639434814,
      "learning_rate": 1.4318617193337908e-05,
      "loss": 0.4635,
      "num_input_tokens_seen": 16874120,
      "step": 25815
    },
    {
      "epoch": 13.532494758909854,
      "grad_norm": 0.3428225815296173,
      "learning_rate": 1.4308280219231812e-05,
      "loss": 0.4548,
      "num_input_tokens_seen": 16878088,
      "step": 25820
    },
    {
      "epoch": 13.535115303983229,
      "grad_norm": 0.4381956160068512,
      "learning_rate": 1.4297945481711527e-05,
      "loss": 0.4213,
      "num_input_tokens_seen": 16880776,
      "step": 25825
    },
    {
      "epoch": 13.537735849056604,
      "grad_norm": 0.3933109641075134,
      "learning_rate": 1.4287612982938946e-05,
      "loss": 0.3923,
      "num_input_tokens_seen": 16884360,
      "step": 25830
    },
    {
      "epoch": 13.54035639412998,
      "grad_norm": 0.415723592042923,
      "learning_rate": 1.4277282725075523e-05,
      "loss": 0.6842,
      "num_input_tokens_seen": 16887080,
      "step": 25835
    },
    {
      "epoch": 13.542976939203355,
      "grad_norm": 0.4612003266811371,
      "learning_rate": 1.426695471028221e-05,
      "loss": 0.4027,
      "num_input_tokens_seen": 16890536,
      "step": 25840
    },
    {
      "epoch": 13.54559748427673,
      "grad_norm": 0.34877458214759827,
      "learning_rate": 1.4256628940719524e-05,
      "loss": 0.3589,
      "num_input_tokens_seen": 16893992,
      "step": 25845
    },
    {
      "epoch": 13.548218029350105,
      "grad_norm": 0.48250359296798706,
      "learning_rate": 1.42463054185475e-05,
      "loss": 0.446,
      "num_input_tokens_seen": 16896904,
      "step": 25850
    },
    {
      "epoch": 13.55083857442348,
      "grad_norm": 0.5445658564567566,
      "learning_rate": 1.4235984145925705e-05,
      "loss": 0.4117,
      "num_input_tokens_seen": 16899880,
      "step": 25855
    },
    {
      "epoch": 13.553459119496855,
      "grad_norm": 0.4728463292121887,
      "learning_rate": 1.4225665125013224e-05,
      "loss": 0.3417,
      "num_input_tokens_seen": 16903528,
      "step": 25860
    },
    {
      "epoch": 13.55607966457023,
      "grad_norm": 0.6890825629234314,
      "learning_rate": 1.4215348357968669e-05,
      "loss": 0.3831,
      "num_input_tokens_seen": 16905736,
      "step": 25865
    },
    {
      "epoch": 13.558700209643606,
      "grad_norm": 0.4297698140144348,
      "learning_rate": 1.42050338469502e-05,
      "loss": 0.3887,
      "num_input_tokens_seen": 16909256,
      "step": 25870
    },
    {
      "epoch": 13.56132075471698,
      "grad_norm": 0.44547039270401,
      "learning_rate": 1.4194721594115495e-05,
      "loss": 0.404,
      "num_input_tokens_seen": 16911976,
      "step": 25875
    },
    {
      "epoch": 13.563941299790356,
      "grad_norm": 0.47869443893432617,
      "learning_rate": 1.4184411601621769e-05,
      "loss": 0.3294,
      "num_input_tokens_seen": 16915592,
      "step": 25880
    },
    {
      "epoch": 13.566561844863731,
      "grad_norm": 0.29397058486938477,
      "learning_rate": 1.4174103871625738e-05,
      "loss": 0.4416,
      "num_input_tokens_seen": 16919080,
      "step": 25885
    },
    {
      "epoch": 13.569182389937106,
      "grad_norm": 0.512440025806427,
      "learning_rate": 1.4163798406283674e-05,
      "loss": 0.5478,
      "num_input_tokens_seen": 16922632,
      "step": 25890
    },
    {
      "epoch": 13.571802935010481,
      "grad_norm": 1.174972653388977,
      "learning_rate": 1.415349520775135e-05,
      "loss": 0.5078,
      "num_input_tokens_seen": 16925736,
      "step": 25895
    },
    {
      "epoch": 13.574423480083858,
      "grad_norm": 0.49391764402389526,
      "learning_rate": 1.4143194278184086e-05,
      "loss": 0.3521,
      "num_input_tokens_seen": 16929448,
      "step": 25900
    },
    {
      "epoch": 13.577044025157234,
      "grad_norm": 0.5104974508285522,
      "learning_rate": 1.4132895619736724e-05,
      "loss": 0.5268,
      "num_input_tokens_seen": 16932584,
      "step": 25905
    },
    {
      "epoch": 13.579664570230609,
      "grad_norm": 0.5136038064956665,
      "learning_rate": 1.4122599234563613e-05,
      "loss": 0.6961,
      "num_input_tokens_seen": 16936072,
      "step": 25910
    },
    {
      "epoch": 13.582285115303984,
      "grad_norm": 0.6598944067955017,
      "learning_rate": 1.4112305124818647e-05,
      "loss": 0.3381,
      "num_input_tokens_seen": 16939464,
      "step": 25915
    },
    {
      "epoch": 13.584905660377359,
      "grad_norm": 0.4474201798439026,
      "learning_rate": 1.4102013292655242e-05,
      "loss": 0.4337,
      "num_input_tokens_seen": 16943912,
      "step": 25920
    },
    {
      "epoch": 13.587526205450734,
      "grad_norm": 0.33597999811172485,
      "learning_rate": 1.409172374022632e-05,
      "loss": 0.5172,
      "num_input_tokens_seen": 16947464,
      "step": 25925
    },
    {
      "epoch": 13.59014675052411,
      "grad_norm": 0.49674898386001587,
      "learning_rate": 1.4081436469684339e-05,
      "loss": 0.3972,
      "num_input_tokens_seen": 16950504,
      "step": 25930
    },
    {
      "epoch": 13.592767295597485,
      "grad_norm": 0.2853606939315796,
      "learning_rate": 1.4071151483181294e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 16953672,
      "step": 25935
    },
    {
      "epoch": 13.59538784067086,
      "grad_norm": 0.5518217086791992,
      "learning_rate": 1.4060868782868669e-05,
      "loss": 0.4099,
      "num_input_tokens_seen": 16956520,
      "step": 25940
    },
    {
      "epoch": 13.598008385744235,
      "grad_norm": 0.6288840770721436,
      "learning_rate": 1.405058837089749e-05,
      "loss": 0.4248,
      "num_input_tokens_seen": 16958888,
      "step": 25945
    },
    {
      "epoch": 13.60062893081761,
      "grad_norm": 0.6835563778877258,
      "learning_rate": 1.404031024941832e-05,
      "loss": 0.4203,
      "num_input_tokens_seen": 16962216,
      "step": 25950
    },
    {
      "epoch": 13.603249475890985,
      "grad_norm": 0.43982642889022827,
      "learning_rate": 1.40300344205812e-05,
      "loss": 0.5561,
      "num_input_tokens_seen": 16966152,
      "step": 25955
    },
    {
      "epoch": 13.60587002096436,
      "grad_norm": 0.44895288348197937,
      "learning_rate": 1.4019760886535738e-05,
      "loss": 0.4553,
      "num_input_tokens_seen": 16969096,
      "step": 25960
    },
    {
      "epoch": 13.608490566037736,
      "grad_norm": 0.28296858072280884,
      "learning_rate": 1.4009489649431018e-05,
      "loss": 0.333,
      "num_input_tokens_seen": 16971592,
      "step": 25965
    },
    {
      "epoch": 13.61111111111111,
      "grad_norm": 0.4726634621620178,
      "learning_rate": 1.3999220711415675e-05,
      "loss": 0.4046,
      "num_input_tokens_seen": 16974056,
      "step": 25970
    },
    {
      "epoch": 13.613731656184486,
      "grad_norm": 0.735377848148346,
      "learning_rate": 1.3988954074637867e-05,
      "loss": 0.3817,
      "num_input_tokens_seen": 16976008,
      "step": 25975
    },
    {
      "epoch": 13.616352201257861,
      "grad_norm": 0.40669307112693787,
      "learning_rate": 1.3978689741245238e-05,
      "loss": 0.4707,
      "num_input_tokens_seen": 16979080,
      "step": 25980
    },
    {
      "epoch": 13.618972746331236,
      "grad_norm": 0.4414253234863281,
      "learning_rate": 1.3968427713384986e-05,
      "loss": 0.5627,
      "num_input_tokens_seen": 16982216,
      "step": 25985
    },
    {
      "epoch": 13.621593291404611,
      "grad_norm": 1.0662508010864258,
      "learning_rate": 1.3958167993203791e-05,
      "loss": 0.4985,
      "num_input_tokens_seen": 16985288,
      "step": 25990
    },
    {
      "epoch": 13.624213836477988,
      "grad_norm": 0.4638270437717438,
      "learning_rate": 1.3947910582847879e-05,
      "loss": 0.3845,
      "num_input_tokens_seen": 16988264,
      "step": 25995
    },
    {
      "epoch": 13.626834381551364,
      "grad_norm": 0.9678370356559753,
      "learning_rate": 1.3937655484462985e-05,
      "loss": 0.4051,
      "num_input_tokens_seen": 16990920,
      "step": 26000
    },
    {
      "epoch": 13.629454926624739,
      "grad_norm": 0.36993205547332764,
      "learning_rate": 1.392740270019437e-05,
      "loss": 0.4338,
      "num_input_tokens_seen": 16993768,
      "step": 26005
    },
    {
      "epoch": 13.632075471698114,
      "grad_norm": 0.5151022672653198,
      "learning_rate": 1.3917152232186776e-05,
      "loss": 0.4554,
      "num_input_tokens_seen": 16996584,
      "step": 26010
    },
    {
      "epoch": 13.634696016771489,
      "grad_norm": 0.4278092384338379,
      "learning_rate": 1.3906904082584504e-05,
      "loss": 0.4799,
      "num_input_tokens_seen": 17000040,
      "step": 26015
    },
    {
      "epoch": 13.637316561844864,
      "grad_norm": 0.5081522464752197,
      "learning_rate": 1.3896658253531334e-05,
      "loss": 0.3648,
      "num_input_tokens_seen": 17002696,
      "step": 26020
    },
    {
      "epoch": 13.63993710691824,
      "grad_norm": 0.29145944118499756,
      "learning_rate": 1.3886414747170581e-05,
      "loss": 0.4888,
      "num_input_tokens_seen": 17005992,
      "step": 26025
    },
    {
      "epoch": 13.642557651991615,
      "grad_norm": 0.4864349365234375,
      "learning_rate": 1.3876173565645072e-05,
      "loss": 0.35,
      "num_input_tokens_seen": 17008904,
      "step": 26030
    },
    {
      "epoch": 13.64517819706499,
      "grad_norm": 1.110539197921753,
      "learning_rate": 1.3865934711097156e-05,
      "loss": 0.3633,
      "num_input_tokens_seen": 17011880,
      "step": 26035
    },
    {
      "epoch": 13.647798742138365,
      "grad_norm": 0.37458086013793945,
      "learning_rate": 1.3855698185668675e-05,
      "loss": 0.4095,
      "num_input_tokens_seen": 17015208,
      "step": 26040
    },
    {
      "epoch": 13.65041928721174,
      "grad_norm": 0.3603766858577728,
      "learning_rate": 1.3845463991500985e-05,
      "loss": 0.5278,
      "num_input_tokens_seen": 17018664,
      "step": 26045
    },
    {
      "epoch": 13.653039832285115,
      "grad_norm": 0.432281494140625,
      "learning_rate": 1.3835232130734968e-05,
      "loss": 0.5099,
      "num_input_tokens_seen": 17022280,
      "step": 26050
    },
    {
      "epoch": 13.65566037735849,
      "grad_norm": 0.768225908279419,
      "learning_rate": 1.3825002605511017e-05,
      "loss": 0.4918,
      "num_input_tokens_seen": 17026344,
      "step": 26055
    },
    {
      "epoch": 13.658280922431866,
      "grad_norm": 0.3344791531562805,
      "learning_rate": 1.3814775417969037e-05,
      "loss": 0.3856,
      "num_input_tokens_seen": 17030024,
      "step": 26060
    },
    {
      "epoch": 13.66090146750524,
      "grad_norm": 0.4142833948135376,
      "learning_rate": 1.3804550570248431e-05,
      "loss": 0.5429,
      "num_input_tokens_seen": 17033992,
      "step": 26065
    },
    {
      "epoch": 13.663522012578616,
      "grad_norm": 0.31886032223701477,
      "learning_rate": 1.3794328064488116e-05,
      "loss": 0.3795,
      "num_input_tokens_seen": 17040936,
      "step": 26070
    },
    {
      "epoch": 13.666142557651991,
      "grad_norm": 0.4563676118850708,
      "learning_rate": 1.3784107902826523e-05,
      "loss": 0.3588,
      "num_input_tokens_seen": 17043432,
      "step": 26075
    },
    {
      "epoch": 13.668763102725366,
      "grad_norm": 1.0168540477752686,
      "learning_rate": 1.3773890087401603e-05,
      "loss": 0.3584,
      "num_input_tokens_seen": 17046184,
      "step": 26080
    },
    {
      "epoch": 13.671383647798741,
      "grad_norm": 0.4808821678161621,
      "learning_rate": 1.3763674620350802e-05,
      "loss": 0.3829,
      "num_input_tokens_seen": 17049160,
      "step": 26085
    },
    {
      "epoch": 13.674004192872118,
      "grad_norm": 0.5687874555587769,
      "learning_rate": 1.3753461503811089e-05,
      "loss": 0.3908,
      "num_input_tokens_seen": 17052584,
      "step": 26090
    },
    {
      "epoch": 13.676624737945493,
      "grad_norm": 0.33939969539642334,
      "learning_rate": 1.3743250739918922e-05,
      "loss": 0.4777,
      "num_input_tokens_seen": 17055624,
      "step": 26095
    },
    {
      "epoch": 13.679245283018869,
      "grad_norm": 0.6348689794540405,
      "learning_rate": 1.3733042330810265e-05,
      "loss": 0.5207,
      "num_input_tokens_seen": 17058280,
      "step": 26100
    },
    {
      "epoch": 13.681865828092244,
      "grad_norm": 0.3372742235660553,
      "learning_rate": 1.3722836278620616e-05,
      "loss": 0.3267,
      "num_input_tokens_seen": 17061256,
      "step": 26105
    },
    {
      "epoch": 13.684486373165619,
      "grad_norm": 0.8110164403915405,
      "learning_rate": 1.3712632585484957e-05,
      "loss": 0.3864,
      "num_input_tokens_seen": 17065160,
      "step": 26110
    },
    {
      "epoch": 13.687106918238994,
      "grad_norm": 0.6963812708854675,
      "learning_rate": 1.37024312535378e-05,
      "loss": 0.526,
      "num_input_tokens_seen": 17067816,
      "step": 26115
    },
    {
      "epoch": 13.68972746331237,
      "grad_norm": 0.277777761220932,
      "learning_rate": 1.3692232284913133e-05,
      "loss": 0.3816,
      "num_input_tokens_seen": 17070952,
      "step": 26120
    },
    {
      "epoch": 13.692348008385745,
      "grad_norm": 0.4321741759777069,
      "learning_rate": 1.3682035681744459e-05,
      "loss": 0.427,
      "num_input_tokens_seen": 17073928,
      "step": 26125
    },
    {
      "epoch": 13.69496855345912,
      "grad_norm": 0.3669869005680084,
      "learning_rate": 1.3671841446164795e-05,
      "loss": 0.4296,
      "num_input_tokens_seen": 17077192,
      "step": 26130
    },
    {
      "epoch": 13.697589098532495,
      "grad_norm": 0.6105644702911377,
      "learning_rate": 1.3661649580306669e-05,
      "loss": 0.4302,
      "num_input_tokens_seen": 17080360,
      "step": 26135
    },
    {
      "epoch": 13.70020964360587,
      "grad_norm": 0.4788050353527069,
      "learning_rate": 1.3651460086302087e-05,
      "loss": 0.4794,
      "num_input_tokens_seen": 17083304,
      "step": 26140
    },
    {
      "epoch": 13.702830188679245,
      "grad_norm": 0.41504982113838196,
      "learning_rate": 1.3641272966282609e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 17086952,
      "step": 26145
    },
    {
      "epoch": 13.70545073375262,
      "grad_norm": 0.3586510419845581,
      "learning_rate": 1.3631088222379218e-05,
      "loss": 0.4168,
      "num_input_tokens_seen": 17091656,
      "step": 26150
    },
    {
      "epoch": 13.708071278825996,
      "grad_norm": 0.4434710144996643,
      "learning_rate": 1.3620905856722468e-05,
      "loss": 0.513,
      "num_input_tokens_seen": 17094888,
      "step": 26155
    },
    {
      "epoch": 13.71069182389937,
      "grad_norm": 0.5444818139076233,
      "learning_rate": 1.3610725871442392e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 17098856,
      "step": 26160
    },
    {
      "epoch": 13.713312368972746,
      "grad_norm": 0.40290364623069763,
      "learning_rate": 1.3600548268668532e-05,
      "loss": 0.3451,
      "num_input_tokens_seen": 17101992,
      "step": 26165
    },
    {
      "epoch": 13.715932914046121,
      "grad_norm": 0.5011420845985413,
      "learning_rate": 1.3590373050529927e-05,
      "loss": 0.5828,
      "num_input_tokens_seen": 17104520,
      "step": 26170
    },
    {
      "epoch": 13.718553459119496,
      "grad_norm": 0.5544834733009338,
      "learning_rate": 1.3580200219155116e-05,
      "loss": 0.42,
      "num_input_tokens_seen": 17107944,
      "step": 26175
    },
    {
      "epoch": 13.721174004192871,
      "grad_norm": 0.5132957100868225,
      "learning_rate": 1.3570029776672122e-05,
      "loss": 0.4698,
      "num_input_tokens_seen": 17111016,
      "step": 26180
    },
    {
      "epoch": 13.723794549266248,
      "grad_norm": 0.5179880857467651,
      "learning_rate": 1.35598617252085e-05,
      "loss": 0.4739,
      "num_input_tokens_seen": 17113544,
      "step": 26185
    },
    {
      "epoch": 13.726415094339622,
      "grad_norm": 0.5967763066291809,
      "learning_rate": 1.354969606689129e-05,
      "loss": 0.3545,
      "num_input_tokens_seen": 17116168,
      "step": 26190
    },
    {
      "epoch": 13.729035639412999,
      "grad_norm": 0.4175252318382263,
      "learning_rate": 1.353953280384704e-05,
      "loss": 0.4136,
      "num_input_tokens_seen": 17120264,
      "step": 26195
    },
    {
      "epoch": 13.731656184486374,
      "grad_norm": 0.9483580589294434,
      "learning_rate": 1.352937193820177e-05,
      "loss": 0.484,
      "num_input_tokens_seen": 17123080,
      "step": 26200
    },
    {
      "epoch": 13.734276729559749,
      "grad_norm": 0.5822458267211914,
      "learning_rate": 1.3519213472081038e-05,
      "loss": 0.4189,
      "num_input_tokens_seen": 17125960,
      "step": 26205
    },
    {
      "epoch": 13.736897274633124,
      "grad_norm": 0.48621463775634766,
      "learning_rate": 1.3509057407609854e-05,
      "loss": 0.5006,
      "num_input_tokens_seen": 17129544,
      "step": 26210
    },
    {
      "epoch": 13.7395178197065,
      "grad_norm": 0.6802675127983093,
      "learning_rate": 1.3498903746912769e-05,
      "loss": 0.4854,
      "num_input_tokens_seen": 17133384,
      "step": 26215
    },
    {
      "epoch": 13.742138364779874,
      "grad_norm": 0.677187442779541,
      "learning_rate": 1.3488752492113807e-05,
      "loss": 0.5453,
      "num_input_tokens_seen": 17136008,
      "step": 26220
    },
    {
      "epoch": 13.74475890985325,
      "grad_norm": 0.7047768235206604,
      "learning_rate": 1.3478603645336508e-05,
      "loss": 0.5367,
      "num_input_tokens_seen": 17139144,
      "step": 26225
    },
    {
      "epoch": 13.747379454926625,
      "grad_norm": 0.4795970618724823,
      "learning_rate": 1.3468457208703877e-05,
      "loss": 0.3159,
      "num_input_tokens_seen": 17142280,
      "step": 26230
    },
    {
      "epoch": 13.75,
      "grad_norm": 0.31006646156311035,
      "learning_rate": 1.3458313184338437e-05,
      "loss": 0.4345,
      "num_input_tokens_seen": 17146664,
      "step": 26235
    },
    {
      "epoch": 13.752620545073375,
      "grad_norm": 0.34344133734703064,
      "learning_rate": 1.3448171574362201e-05,
      "loss": 0.4924,
      "num_input_tokens_seen": 17149384,
      "step": 26240
    },
    {
      "epoch": 13.75524109014675,
      "grad_norm": 0.5505586266517639,
      "learning_rate": 1.3438032380896681e-05,
      "loss": 0.5258,
      "num_input_tokens_seen": 17152328,
      "step": 26245
    },
    {
      "epoch": 13.757861635220126,
      "grad_norm": 0.3638513386249542,
      "learning_rate": 1.342789560606289e-05,
      "loss": 0.4125,
      "num_input_tokens_seen": 17155624,
      "step": 26250
    },
    {
      "epoch": 13.7604821802935,
      "grad_norm": 0.43442502617836,
      "learning_rate": 1.341776125198131e-05,
      "loss": 0.4039,
      "num_input_tokens_seen": 17158376,
      "step": 26255
    },
    {
      "epoch": 13.763102725366876,
      "grad_norm": 0.3688223958015442,
      "learning_rate": 1.3407629320771948e-05,
      "loss": 0.3934,
      "num_input_tokens_seen": 17161320,
      "step": 26260
    },
    {
      "epoch": 13.765723270440251,
      "grad_norm": 0.4342171549797058,
      "learning_rate": 1.3397499814554266e-05,
      "loss": 0.5047,
      "num_input_tokens_seen": 17165288,
      "step": 26265
    },
    {
      "epoch": 13.768343815513626,
      "grad_norm": 0.29157036542892456,
      "learning_rate": 1.3387372735447256e-05,
      "loss": 0.4056,
      "num_input_tokens_seen": 17168456,
      "step": 26270
    },
    {
      "epoch": 13.770964360587001,
      "grad_norm": 0.46070852875709534,
      "learning_rate": 1.3377248085569399e-05,
      "loss": 0.4332,
      "num_input_tokens_seen": 17171496,
      "step": 26275
    },
    {
      "epoch": 13.773584905660378,
      "grad_norm": 0.5342812538146973,
      "learning_rate": 1.336712586703863e-05,
      "loss": 0.4014,
      "num_input_tokens_seen": 17174824,
      "step": 26280
    },
    {
      "epoch": 13.776205450733752,
      "grad_norm": 0.5901153683662415,
      "learning_rate": 1.3357006081972414e-05,
      "loss": 0.5714,
      "num_input_tokens_seen": 17177064,
      "step": 26285
    },
    {
      "epoch": 13.778825995807129,
      "grad_norm": 0.7398769855499268,
      "learning_rate": 1.3346888732487706e-05,
      "loss": 0.3886,
      "num_input_tokens_seen": 17179752,
      "step": 26290
    },
    {
      "epoch": 13.781446540880504,
      "grad_norm": 0.26994213461875916,
      "learning_rate": 1.3336773820700921e-05,
      "loss": 0.5165,
      "num_input_tokens_seen": 17184232,
      "step": 26295
    },
    {
      "epoch": 13.784067085953879,
      "grad_norm": 0.5570823550224304,
      "learning_rate": 1.3326661348727992e-05,
      "loss": 0.3614,
      "num_input_tokens_seen": 17187624,
      "step": 26300
    },
    {
      "epoch": 13.786687631027254,
      "grad_norm": 0.6898630857467651,
      "learning_rate": 1.3316551318684342e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 17190088,
      "step": 26305
    },
    {
      "epoch": 13.78930817610063,
      "grad_norm": 0.33473390340805054,
      "learning_rate": 1.3306443732684854e-05,
      "loss": 0.4538,
      "num_input_tokens_seen": 17194216,
      "step": 26310
    },
    {
      "epoch": 13.791928721174004,
      "grad_norm": 0.4441421329975128,
      "learning_rate": 1.3296338592843929e-05,
      "loss": 0.4562,
      "num_input_tokens_seen": 17197544,
      "step": 26315
    },
    {
      "epoch": 13.79454926624738,
      "grad_norm": 0.7373020648956299,
      "learning_rate": 1.3286235901275457e-05,
      "loss": 0.5289,
      "num_input_tokens_seen": 17200488,
      "step": 26320
    },
    {
      "epoch": 13.797169811320755,
      "grad_norm": 0.9440532922744751,
      "learning_rate": 1.327613566009279e-05,
      "loss": 0.4609,
      "num_input_tokens_seen": 17202824,
      "step": 26325
    },
    {
      "epoch": 13.79979035639413,
      "grad_norm": 0.7122911214828491,
      "learning_rate": 1.3266037871408798e-05,
      "loss": 0.3355,
      "num_input_tokens_seen": 17206632,
      "step": 26330
    },
    {
      "epoch": 13.802410901467505,
      "grad_norm": 0.13315458595752716,
      "learning_rate": 1.3255942537335805e-05,
      "loss": 0.3035,
      "num_input_tokens_seen": 17212648,
      "step": 26335
    },
    {
      "epoch": 13.80503144654088,
      "grad_norm": 0.2983167767524719,
      "learning_rate": 1.3245849659985652e-05,
      "loss": 0.429,
      "num_input_tokens_seen": 17215176,
      "step": 26340
    },
    {
      "epoch": 13.807651991614255,
      "grad_norm": 0.7585946321487427,
      "learning_rate": 1.3235759241469658e-05,
      "loss": 0.4441,
      "num_input_tokens_seen": 17217672,
      "step": 26345
    },
    {
      "epoch": 13.81027253668763,
      "grad_norm": 0.2440846562385559,
      "learning_rate": 1.322567128389861e-05,
      "loss": 0.3584,
      "num_input_tokens_seen": 17221096,
      "step": 26350
    },
    {
      "epoch": 13.812893081761006,
      "grad_norm": 0.5687814354896545,
      "learning_rate": 1.3215585789382812e-05,
      "loss": 0.3858,
      "num_input_tokens_seen": 17223784,
      "step": 26355
    },
    {
      "epoch": 13.815513626834381,
      "grad_norm": 0.4288937747478485,
      "learning_rate": 1.320550276003201e-05,
      "loss": 0.4969,
      "num_input_tokens_seen": 17227496,
      "step": 26360
    },
    {
      "epoch": 13.818134171907756,
      "grad_norm": 0.5406330823898315,
      "learning_rate": 1.3195422197955473e-05,
      "loss": 0.4497,
      "num_input_tokens_seen": 17230728,
      "step": 26365
    },
    {
      "epoch": 13.820754716981131,
      "grad_norm": 0.432322233915329,
      "learning_rate": 1.3185344105261937e-05,
      "loss": 0.4682,
      "num_input_tokens_seen": 17233800,
      "step": 26370
    },
    {
      "epoch": 13.823375262054507,
      "grad_norm": 0.5212472677230835,
      "learning_rate": 1.3175268484059639e-05,
      "loss": 0.4214,
      "num_input_tokens_seen": 17236744,
      "step": 26375
    },
    {
      "epoch": 13.825995807127882,
      "grad_norm": 0.3640708923339844,
      "learning_rate": 1.316519533645626e-05,
      "loss": 0.4024,
      "num_input_tokens_seen": 17241224,
      "step": 26380
    },
    {
      "epoch": 13.828616352201259,
      "grad_norm": 0.6828746795654297,
      "learning_rate": 1.3155124664559004e-05,
      "loss": 0.4735,
      "num_input_tokens_seen": 17244808,
      "step": 26385
    },
    {
      "epoch": 13.831236897274634,
      "grad_norm": 0.45623454451560974,
      "learning_rate": 1.3145056470474525e-05,
      "loss": 0.3845,
      "num_input_tokens_seen": 17248168,
      "step": 26390
    },
    {
      "epoch": 13.833857442348009,
      "grad_norm": 0.33289405703544617,
      "learning_rate": 1.313499075630899e-05,
      "loss": 0.4755,
      "num_input_tokens_seen": 17252136,
      "step": 26395
    },
    {
      "epoch": 13.836477987421384,
      "grad_norm": 0.360539048910141,
      "learning_rate": 1.312492752416802e-05,
      "loss": 0.4968,
      "num_input_tokens_seen": 17255880,
      "step": 26400
    },
    {
      "epoch": 13.83909853249476,
      "grad_norm": 0.6777201890945435,
      "learning_rate": 1.3114866776156748e-05,
      "loss": 0.642,
      "num_input_tokens_seen": 17258728,
      "step": 26405
    },
    {
      "epoch": 13.841719077568134,
      "grad_norm": 0.3542881906032562,
      "learning_rate": 1.3104808514379748e-05,
      "loss": 0.4545,
      "num_input_tokens_seen": 17261992,
      "step": 26410
    },
    {
      "epoch": 13.84433962264151,
      "grad_norm": 0.40043702721595764,
      "learning_rate": 1.3094752740941094e-05,
      "loss": 0.3673,
      "num_input_tokens_seen": 17265896,
      "step": 26415
    },
    {
      "epoch": 13.846960167714885,
      "grad_norm": 0.3688475787639618,
      "learning_rate": 1.3084699457944344e-05,
      "loss": 0.466,
      "num_input_tokens_seen": 17269032,
      "step": 26420
    },
    {
      "epoch": 13.84958071278826,
      "grad_norm": 0.4945465624332428,
      "learning_rate": 1.3074648667492528e-05,
      "loss": 0.4612,
      "num_input_tokens_seen": 17272456,
      "step": 26425
    },
    {
      "epoch": 13.852201257861635,
      "grad_norm": 0.4330224394798279,
      "learning_rate": 1.3064600371688171e-05,
      "loss": 0.464,
      "num_input_tokens_seen": 17275496,
      "step": 26430
    },
    {
      "epoch": 13.85482180293501,
      "grad_norm": 0.6671005487442017,
      "learning_rate": 1.3054554572633249e-05,
      "loss": 0.3378,
      "num_input_tokens_seen": 17277672,
      "step": 26435
    },
    {
      "epoch": 13.857442348008385,
      "grad_norm": 0.26997247338294983,
      "learning_rate": 1.3044511272429222e-05,
      "loss": 0.461,
      "num_input_tokens_seen": 17281768,
      "step": 26440
    },
    {
      "epoch": 13.86006289308176,
      "grad_norm": 0.40058213472366333,
      "learning_rate": 1.3034470473177038e-05,
      "loss": 0.4687,
      "num_input_tokens_seen": 17284616,
      "step": 26445
    },
    {
      "epoch": 13.862683438155136,
      "grad_norm": 0.6738510727882385,
      "learning_rate": 1.3024432176977125e-05,
      "loss": 0.5463,
      "num_input_tokens_seen": 17287848,
      "step": 26450
    },
    {
      "epoch": 13.865303983228511,
      "grad_norm": 0.5142555832862854,
      "learning_rate": 1.3014396385929372e-05,
      "loss": 0.3812,
      "num_input_tokens_seen": 17294312,
      "step": 26455
    },
    {
      "epoch": 13.867924528301886,
      "grad_norm": 0.6603137850761414,
      "learning_rate": 1.3004363102133165e-05,
      "loss": 0.5693,
      "num_input_tokens_seen": 17297672,
      "step": 26460
    },
    {
      "epoch": 13.870545073375261,
      "grad_norm": 0.47069665789604187,
      "learning_rate": 1.2994332327687348e-05,
      "loss": 0.4482,
      "num_input_tokens_seen": 17301032,
      "step": 26465
    },
    {
      "epoch": 13.873165618448636,
      "grad_norm": 0.3165799677371979,
      "learning_rate": 1.2984304064690228e-05,
      "loss": 0.4065,
      "num_input_tokens_seen": 17304200,
      "step": 26470
    },
    {
      "epoch": 13.875786163522012,
      "grad_norm": 0.6877913475036621,
      "learning_rate": 1.2974278315239619e-05,
      "loss": 0.4726,
      "num_input_tokens_seen": 17307208,
      "step": 26475
    },
    {
      "epoch": 13.878406708595389,
      "grad_norm": 0.33874696493148804,
      "learning_rate": 1.2964255081432786e-05,
      "loss": 0.3871,
      "num_input_tokens_seen": 17310056,
      "step": 26480
    },
    {
      "epoch": 13.881027253668764,
      "grad_norm": 0.2584666311740875,
      "learning_rate": 1.2954234365366488e-05,
      "loss": 0.4153,
      "num_input_tokens_seen": 17313960,
      "step": 26485
    },
    {
      "epoch": 13.883647798742139,
      "grad_norm": 0.33595409989356995,
      "learning_rate": 1.2944216169136936e-05,
      "loss": 0.3761,
      "num_input_tokens_seen": 17316712,
      "step": 26490
    },
    {
      "epoch": 13.886268343815514,
      "grad_norm": 0.394318163394928,
      "learning_rate": 1.2934200494839815e-05,
      "loss": 0.3921,
      "num_input_tokens_seen": 17321096,
      "step": 26495
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 0.7174528241157532,
      "learning_rate": 1.2924187344570293e-05,
      "loss": 0.571,
      "num_input_tokens_seen": 17324168,
      "step": 26500
    },
    {
      "epoch": 13.891509433962264,
      "grad_norm": 0.54697185754776,
      "learning_rate": 1.2914176720423011e-05,
      "loss": 0.3992,
      "num_input_tokens_seen": 17327016,
      "step": 26505
    },
    {
      "epoch": 13.89412997903564,
      "grad_norm": 0.67316073179245,
      "learning_rate": 1.2904168624492088e-05,
      "loss": 0.5166,
      "num_input_tokens_seen": 17330088,
      "step": 26510
    },
    {
      "epoch": 13.896750524109015,
      "grad_norm": 0.6362142562866211,
      "learning_rate": 1.2894163058871092e-05,
      "loss": 0.3776,
      "num_input_tokens_seen": 17332328,
      "step": 26515
    },
    {
      "epoch": 13.89937106918239,
      "grad_norm": 0.6169623732566833,
      "learning_rate": 1.2884160025653064e-05,
      "loss": 0.4134,
      "num_input_tokens_seen": 17336584,
      "step": 26520
    },
    {
      "epoch": 13.901991614255765,
      "grad_norm": 0.5770593285560608,
      "learning_rate": 1.2874159526930532e-05,
      "loss": 0.4671,
      "num_input_tokens_seen": 17339240,
      "step": 26525
    },
    {
      "epoch": 13.90461215932914,
      "grad_norm": 0.39283499121665955,
      "learning_rate": 1.286416156479549e-05,
      "loss": 0.4442,
      "num_input_tokens_seen": 17343688,
      "step": 26530
    },
    {
      "epoch": 13.907232704402515,
      "grad_norm": 0.3584994673728943,
      "learning_rate": 1.2854166141339394e-05,
      "loss": 0.3823,
      "num_input_tokens_seen": 17347336,
      "step": 26535
    },
    {
      "epoch": 13.90985324947589,
      "grad_norm": 0.3839951157569885,
      "learning_rate": 1.2844173258653183e-05,
      "loss": 0.4143,
      "num_input_tokens_seen": 17350952,
      "step": 26540
    },
    {
      "epoch": 13.912473794549266,
      "grad_norm": 0.5343257784843445,
      "learning_rate": 1.2834182918827247e-05,
      "loss": 0.5346,
      "num_input_tokens_seen": 17354376,
      "step": 26545
    },
    {
      "epoch": 13.915094339622641,
      "grad_norm": 0.9644942879676819,
      "learning_rate": 1.282419512395144e-05,
      "loss": 0.5169,
      "num_input_tokens_seen": 17358216,
      "step": 26550
    },
    {
      "epoch": 13.917714884696016,
      "grad_norm": 0.6425473093986511,
      "learning_rate": 1.2814209876115102e-05,
      "loss": 0.4278,
      "num_input_tokens_seen": 17361224,
      "step": 26555
    },
    {
      "epoch": 13.920335429769391,
      "grad_norm": 0.42200714349746704,
      "learning_rate": 1.2804227177407038e-05,
      "loss": 0.4388,
      "num_input_tokens_seen": 17364488,
      "step": 26560
    },
    {
      "epoch": 13.922955974842766,
      "grad_norm": 0.6695373058319092,
      "learning_rate": 1.279424702991552e-05,
      "loss": 0.4503,
      "num_input_tokens_seen": 17368936,
      "step": 26565
    },
    {
      "epoch": 13.925576519916142,
      "grad_norm": 0.5635443329811096,
      "learning_rate": 1.2784269435728264e-05,
      "loss": 0.483,
      "num_input_tokens_seen": 17371912,
      "step": 26570
    },
    {
      "epoch": 13.928197064989519,
      "grad_norm": 0.5498172640800476,
      "learning_rate": 1.2774294396932488e-05,
      "loss": 0.3826,
      "num_input_tokens_seen": 17374856,
      "step": 26575
    },
    {
      "epoch": 13.930817610062894,
      "grad_norm": 0.8205374479293823,
      "learning_rate": 1.2764321915614835e-05,
      "loss": 0.3031,
      "num_input_tokens_seen": 17377832,
      "step": 26580
    },
    {
      "epoch": 13.933438155136269,
      "grad_norm": 0.9814174175262451,
      "learning_rate": 1.2754351993861444e-05,
      "loss": 0.4506,
      "num_input_tokens_seen": 17381064,
      "step": 26585
    },
    {
      "epoch": 13.936058700209644,
      "grad_norm": 0.28250905871391296,
      "learning_rate": 1.2744384633757922e-05,
      "loss": 0.3933,
      "num_input_tokens_seen": 17384968,
      "step": 26590
    },
    {
      "epoch": 13.93867924528302,
      "grad_norm": 0.411233514547348,
      "learning_rate": 1.2734419837389306e-05,
      "loss": 0.499,
      "num_input_tokens_seen": 17388104,
      "step": 26595
    },
    {
      "epoch": 13.941299790356394,
      "grad_norm": 0.9391133785247803,
      "learning_rate": 1.2724457606840123e-05,
      "loss": 0.5032,
      "num_input_tokens_seen": 17391880,
      "step": 26600
    },
    {
      "epoch": 13.94392033542977,
      "grad_norm": 0.3713204860687256,
      "learning_rate": 1.2714497944194376e-05,
      "loss": 0.4259,
      "num_input_tokens_seen": 17395240,
      "step": 26605
    },
    {
      "epoch": 13.946540880503145,
      "grad_norm": 0.7056326270103455,
      "learning_rate": 1.2704540851535484e-05,
      "loss": 0.3447,
      "num_input_tokens_seen": 17400680,
      "step": 26610
    },
    {
      "epoch": 13.94916142557652,
      "grad_norm": 0.5775526762008667,
      "learning_rate": 1.2694586330946376e-05,
      "loss": 0.7507,
      "num_input_tokens_seen": 17403080,
      "step": 26615
    },
    {
      "epoch": 13.951781970649895,
      "grad_norm": 0.36612850427627563,
      "learning_rate": 1.2684634384509427e-05,
      "loss": 0.4722,
      "num_input_tokens_seen": 17406600,
      "step": 26620
    },
    {
      "epoch": 13.95440251572327,
      "grad_norm": 0.3658043444156647,
      "learning_rate": 1.2674685014306456e-05,
      "loss": 0.5483,
      "num_input_tokens_seen": 17409320,
      "step": 26625
    },
    {
      "epoch": 13.957023060796645,
      "grad_norm": 0.7857415676116943,
      "learning_rate": 1.2664738222418776e-05,
      "loss": 0.4431,
      "num_input_tokens_seen": 17412616,
      "step": 26630
    },
    {
      "epoch": 13.95964360587002,
      "grad_norm": 0.5341339707374573,
      "learning_rate": 1.2654794010927118e-05,
      "loss": 0.3378,
      "num_input_tokens_seen": 17415912,
      "step": 26635
    },
    {
      "epoch": 13.962264150943396,
      "grad_norm": 0.562197208404541,
      "learning_rate": 1.2644852381911715e-05,
      "loss": 0.479,
      "num_input_tokens_seen": 17419656,
      "step": 26640
    },
    {
      "epoch": 13.964884696016771,
      "grad_norm": 0.597145140171051,
      "learning_rate": 1.2634913337452248e-05,
      "loss": 0.378,
      "num_input_tokens_seen": 17421768,
      "step": 26645
    },
    {
      "epoch": 13.967505241090146,
      "grad_norm": 0.40470418334007263,
      "learning_rate": 1.2624976879627832e-05,
      "loss": 0.3495,
      "num_input_tokens_seen": 17424968,
      "step": 26650
    },
    {
      "epoch": 13.970125786163521,
      "grad_norm": 0.35604196786880493,
      "learning_rate": 1.2615043010517069e-05,
      "loss": 0.3719,
      "num_input_tokens_seen": 17427912,
      "step": 26655
    },
    {
      "epoch": 13.972746331236896,
      "grad_norm": 0.4925672709941864,
      "learning_rate": 1.2605111732198027e-05,
      "loss": 0.3933,
      "num_input_tokens_seen": 17431272,
      "step": 26660
    },
    {
      "epoch": 13.975366876310272,
      "grad_norm": 0.43618613481521606,
      "learning_rate": 1.2595183046748188e-05,
      "loss": 0.4421,
      "num_input_tokens_seen": 17434728,
      "step": 26665
    },
    {
      "epoch": 13.977987421383649,
      "grad_norm": 0.7269288897514343,
      "learning_rate": 1.2585256956244545e-05,
      "loss": 0.354,
      "num_input_tokens_seen": 17438216,
      "step": 26670
    },
    {
      "epoch": 13.980607966457024,
      "grad_norm": 0.40858194231987,
      "learning_rate": 1.2575333462763506e-05,
      "loss": 0.4831,
      "num_input_tokens_seen": 17441384,
      "step": 26675
    },
    {
      "epoch": 13.983228511530399,
      "grad_norm": 0.8338379263877869,
      "learning_rate": 1.2565412568380957e-05,
      "loss": 0.3772,
      "num_input_tokens_seen": 17444168,
      "step": 26680
    },
    {
      "epoch": 13.985849056603774,
      "grad_norm": 0.49293795228004456,
      "learning_rate": 1.2555494275172237e-05,
      "loss": 0.4457,
      "num_input_tokens_seen": 17448168,
      "step": 26685
    },
    {
      "epoch": 13.98846960167715,
      "grad_norm": 0.8332372307777405,
      "learning_rate": 1.2545578585212148e-05,
      "loss": 0.4676,
      "num_input_tokens_seen": 17450824,
      "step": 26690
    },
    {
      "epoch": 13.991090146750524,
      "grad_norm": 0.30719754099845886,
      "learning_rate": 1.2535665500574922e-05,
      "loss": 0.3542,
      "num_input_tokens_seen": 17453832,
      "step": 26695
    },
    {
      "epoch": 13.9937106918239,
      "grad_norm": 0.2023920863866806,
      "learning_rate": 1.2525755023334285e-05,
      "loss": 0.3832,
      "num_input_tokens_seen": 17457704,
      "step": 26700
    },
    {
      "epoch": 13.996331236897275,
      "grad_norm": 0.44527125358581543,
      "learning_rate": 1.251584715556337e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 17460680,
      "step": 26705
    },
    {
      "epoch": 13.99895178197065,
      "grad_norm": 0.39897722005844116,
      "learning_rate": 1.2505941899334805e-05,
      "loss": 0.3655,
      "num_input_tokens_seen": 17464072,
      "step": 26710
    },
    {
      "epoch": 14.0,
      "eval_loss": 0.47587457299232483,
      "eval_runtime": 9.7454,
      "eval_samples_per_second": 87.015,
      "eval_steps_per_second": 21.754,
      "num_input_tokens_seen": 17464792,
      "step": 26712
    },
    {
      "epoch": 14.001572327044025,
      "grad_norm": 0.3420253098011017,
      "learning_rate": 1.2496039256720665e-05,
      "loss": 0.4429,
      "num_input_tokens_seen": 17466744,
      "step": 26715
    },
    {
      "epoch": 14.0041928721174,
      "grad_norm": 0.3725551664829254,
      "learning_rate": 1.2486139229792452e-05,
      "loss": 0.388,
      "num_input_tokens_seen": 17470072,
      "step": 26720
    },
    {
      "epoch": 14.006813417190775,
      "grad_norm": 0.3376474678516388,
      "learning_rate": 1.2476241820621152e-05,
      "loss": 0.3305,
      "num_input_tokens_seen": 17473432,
      "step": 26725
    },
    {
      "epoch": 14.00943396226415,
      "grad_norm": 0.3418564200401306,
      "learning_rate": 1.2466347031277176e-05,
      "loss": 0.3582,
      "num_input_tokens_seen": 17476568,
      "step": 26730
    },
    {
      "epoch": 14.012054507337526,
      "grad_norm": 0.31393975019454956,
      "learning_rate": 1.2456454863830411e-05,
      "loss": 0.4047,
      "num_input_tokens_seen": 17479608,
      "step": 26735
    },
    {
      "epoch": 14.014675052410901,
      "grad_norm": 0.4348782002925873,
      "learning_rate": 1.2446565320350185e-05,
      "loss": 0.477,
      "num_input_tokens_seen": 17482904,
      "step": 26740
    },
    {
      "epoch": 14.017295597484276,
      "grad_norm": 0.4378909468650818,
      "learning_rate": 1.2436678402905283e-05,
      "loss": 0.4673,
      "num_input_tokens_seen": 17485432,
      "step": 26745
    },
    {
      "epoch": 14.019916142557651,
      "grad_norm": 0.6625964641571045,
      "learning_rate": 1.2426794113563933e-05,
      "loss": 0.2944,
      "num_input_tokens_seen": 17487864,
      "step": 26750
    },
    {
      "epoch": 14.022536687631026,
      "grad_norm": 0.5923157334327698,
      "learning_rate": 1.24169124543938e-05,
      "loss": 0.4013,
      "num_input_tokens_seen": 17490424,
      "step": 26755
    },
    {
      "epoch": 14.025157232704403,
      "grad_norm": 0.4637185037136078,
      "learning_rate": 1.2407033427462026e-05,
      "loss": 0.3859,
      "num_input_tokens_seen": 17492888,
      "step": 26760
    },
    {
      "epoch": 14.027777777777779,
      "grad_norm": 0.6734350919723511,
      "learning_rate": 1.2397157034835192e-05,
      "loss": 0.5806,
      "num_input_tokens_seen": 17495448,
      "step": 26765
    },
    {
      "epoch": 14.030398322851154,
      "grad_norm": 0.5144055485725403,
      "learning_rate": 1.2387283278579325e-05,
      "loss": 0.3435,
      "num_input_tokens_seen": 17499160,
      "step": 26770
    },
    {
      "epoch": 14.033018867924529,
      "grad_norm": 0.37525320053100586,
      "learning_rate": 1.237741216075991e-05,
      "loss": 0.4805,
      "num_input_tokens_seen": 17501880,
      "step": 26775
    },
    {
      "epoch": 14.035639412997904,
      "grad_norm": 0.6150497198104858,
      "learning_rate": 1.2367543683441868e-05,
      "loss": 0.5405,
      "num_input_tokens_seen": 17504696,
      "step": 26780
    },
    {
      "epoch": 14.03825995807128,
      "grad_norm": 0.6576557755470276,
      "learning_rate": 1.2357677848689558e-05,
      "loss": 0.4651,
      "num_input_tokens_seen": 17507672,
      "step": 26785
    },
    {
      "epoch": 14.040880503144654,
      "grad_norm": 0.3590565621852875,
      "learning_rate": 1.2347814658566811e-05,
      "loss": 0.3913,
      "num_input_tokens_seen": 17510296,
      "step": 26790
    },
    {
      "epoch": 14.04350104821803,
      "grad_norm": 0.5588021874427795,
      "learning_rate": 1.2337954115136894e-05,
      "loss": 0.4396,
      "num_input_tokens_seen": 17512920,
      "step": 26795
    },
    {
      "epoch": 14.046121593291405,
      "grad_norm": 0.3064824044704437,
      "learning_rate": 1.2328096220462529e-05,
      "loss": 0.4573,
      "num_input_tokens_seen": 17516888,
      "step": 26800
    },
    {
      "epoch": 14.04874213836478,
      "grad_norm": 0.524175226688385,
      "learning_rate": 1.2318240976605865e-05,
      "loss": 0.3877,
      "num_input_tokens_seen": 17520792,
      "step": 26805
    },
    {
      "epoch": 14.051362683438155,
      "grad_norm": 0.5811550617218018,
      "learning_rate": 1.23083883856285e-05,
      "loss": 0.3859,
      "num_input_tokens_seen": 17523896,
      "step": 26810
    },
    {
      "epoch": 14.05398322851153,
      "grad_norm": 0.5319822430610657,
      "learning_rate": 1.229853844959149e-05,
      "loss": 0.45,
      "num_input_tokens_seen": 17527352,
      "step": 26815
    },
    {
      "epoch": 14.056603773584905,
      "grad_norm": 0.4187999963760376,
      "learning_rate": 1.228869117055533e-05,
      "loss": 0.434,
      "num_input_tokens_seen": 17530872,
      "step": 26820
    },
    {
      "epoch": 14.05922431865828,
      "grad_norm": 0.3263239562511444,
      "learning_rate": 1.227884655057997e-05,
      "loss": 0.4799,
      "num_input_tokens_seen": 17533528,
      "step": 26825
    },
    {
      "epoch": 14.061844863731656,
      "grad_norm": 0.34576934576034546,
      "learning_rate": 1.2269004591724777e-05,
      "loss": 0.3156,
      "num_input_tokens_seen": 17537368,
      "step": 26830
    },
    {
      "epoch": 14.064465408805031,
      "grad_norm": 0.5946719646453857,
      "learning_rate": 1.2259165296048575e-05,
      "loss": 0.4418,
      "num_input_tokens_seen": 17541080,
      "step": 26835
    },
    {
      "epoch": 14.067085953878406,
      "grad_norm": 0.4883779287338257,
      "learning_rate": 1.2249328665609638e-05,
      "loss": 0.3632,
      "num_input_tokens_seen": 17545112,
      "step": 26840
    },
    {
      "epoch": 14.069706498951781,
      "grad_norm": 0.7488662600517273,
      "learning_rate": 1.2239494702465676e-05,
      "loss": 0.5085,
      "num_input_tokens_seen": 17547960,
      "step": 26845
    },
    {
      "epoch": 14.072327044025156,
      "grad_norm": 0.26660388708114624,
      "learning_rate": 1.2229663408673842e-05,
      "loss": 0.4836,
      "num_input_tokens_seen": 17551416,
      "step": 26850
    },
    {
      "epoch": 14.074947589098532,
      "grad_norm": 0.4937126636505127,
      "learning_rate": 1.221983478629074e-05,
      "loss": 0.431,
      "num_input_tokens_seen": 17555096,
      "step": 26855
    },
    {
      "epoch": 14.077568134171909,
      "grad_norm": 0.7243624925613403,
      "learning_rate": 1.22100088373724e-05,
      "loss": 0.4847,
      "num_input_tokens_seen": 17559448,
      "step": 26860
    },
    {
      "epoch": 14.080188679245284,
      "grad_norm": 0.4804067611694336,
      "learning_rate": 1.2200185563974283e-05,
      "loss": 0.4132,
      "num_input_tokens_seen": 17562584,
      "step": 26865
    },
    {
      "epoch": 14.082809224318659,
      "grad_norm": 0.33105549216270447,
      "learning_rate": 1.219036496815132e-05,
      "loss": 0.3722,
      "num_input_tokens_seen": 17566776,
      "step": 26870
    },
    {
      "epoch": 14.085429769392034,
      "grad_norm": 0.6077175736427307,
      "learning_rate": 1.2180547051957863e-05,
      "loss": 0.5083,
      "num_input_tokens_seen": 17569368,
      "step": 26875
    },
    {
      "epoch": 14.08805031446541,
      "grad_norm": 0.7316566109657288,
      "learning_rate": 1.2170731817447723e-05,
      "loss": 0.4481,
      "num_input_tokens_seen": 17571928,
      "step": 26880
    },
    {
      "epoch": 14.090670859538784,
      "grad_norm": 0.4634678363800049,
      "learning_rate": 1.216091926667412e-05,
      "loss": 0.4493,
      "num_input_tokens_seen": 17574392,
      "step": 26885
    },
    {
      "epoch": 14.09329140461216,
      "grad_norm": 1.4676977396011353,
      "learning_rate": 1.2151109401689723e-05,
      "loss": 0.3818,
      "num_input_tokens_seen": 17578104,
      "step": 26890
    },
    {
      "epoch": 14.095911949685535,
      "grad_norm": 0.4207656979560852,
      "learning_rate": 1.214130222454665e-05,
      "loss": 0.4088,
      "num_input_tokens_seen": 17581368,
      "step": 26895
    },
    {
      "epoch": 14.09853249475891,
      "grad_norm": 0.30892297625541687,
      "learning_rate": 1.2131497737296452e-05,
      "loss": 0.3448,
      "num_input_tokens_seen": 17584024,
      "step": 26900
    },
    {
      "epoch": 14.101153039832285,
      "grad_norm": 0.6375135183334351,
      "learning_rate": 1.2121695941990124e-05,
      "loss": 0.3628,
      "num_input_tokens_seen": 17586168,
      "step": 26905
    },
    {
      "epoch": 14.10377358490566,
      "grad_norm": 0.5146982073783875,
      "learning_rate": 1.2111896840678075e-05,
      "loss": 0.4115,
      "num_input_tokens_seen": 17591096,
      "step": 26910
    },
    {
      "epoch": 14.106394129979035,
      "grad_norm": 0.7345038652420044,
      "learning_rate": 1.210210043541018e-05,
      "loss": 0.5439,
      "num_input_tokens_seen": 17594264,
      "step": 26915
    },
    {
      "epoch": 14.10901467505241,
      "grad_norm": 0.3792692720890045,
      "learning_rate": 1.2092306728235717e-05,
      "loss": 0.3317,
      "num_input_tokens_seen": 17597944,
      "step": 26920
    },
    {
      "epoch": 14.111635220125786,
      "grad_norm": 0.4713839292526245,
      "learning_rate": 1.2082515721203427e-05,
      "loss": 0.4385,
      "num_input_tokens_seen": 17600728,
      "step": 26925
    },
    {
      "epoch": 14.114255765199161,
      "grad_norm": 0.6962279677391052,
      "learning_rate": 1.2072727416361479e-05,
      "loss": 0.5144,
      "num_input_tokens_seen": 17603576,
      "step": 26930
    },
    {
      "epoch": 14.116876310272536,
      "grad_norm": 0.5967344045639038,
      "learning_rate": 1.2062941815757484e-05,
      "loss": 0.4951,
      "num_input_tokens_seen": 17606392,
      "step": 26935
    },
    {
      "epoch": 14.119496855345911,
      "grad_norm": 0.30986717343330383,
      "learning_rate": 1.2053158921438454e-05,
      "loss": 0.4596,
      "num_input_tokens_seen": 17609944,
      "step": 26940
    },
    {
      "epoch": 14.122117400419286,
      "grad_norm": 0.3944929838180542,
      "learning_rate": 1.2043378735450885e-05,
      "loss": 0.5571,
      "num_input_tokens_seen": 17614424,
      "step": 26945
    },
    {
      "epoch": 14.124737945492662,
      "grad_norm": 0.441382497549057,
      "learning_rate": 1.2033601259840657e-05,
      "loss": 0.449,
      "num_input_tokens_seen": 17617368,
      "step": 26950
    },
    {
      "epoch": 14.127358490566039,
      "grad_norm": 0.6732641458511353,
      "learning_rate": 1.2023826496653118e-05,
      "loss": 0.3702,
      "num_input_tokens_seen": 17620792,
      "step": 26955
    },
    {
      "epoch": 14.129979035639414,
      "grad_norm": 0.5085803866386414,
      "learning_rate": 1.2014054447933044e-05,
      "loss": 0.5008,
      "num_input_tokens_seen": 17624152,
      "step": 26960
    },
    {
      "epoch": 14.132599580712789,
      "grad_norm": 0.44884157180786133,
      "learning_rate": 1.200428511572462e-05,
      "loss": 0.3639,
      "num_input_tokens_seen": 17627064,
      "step": 26965
    },
    {
      "epoch": 14.135220125786164,
      "grad_norm": 0.6533878445625305,
      "learning_rate": 1.1994518502071484e-05,
      "loss": 0.4567,
      "num_input_tokens_seen": 17629720,
      "step": 26970
    },
    {
      "epoch": 14.13784067085954,
      "grad_norm": 0.2764056622982025,
      "learning_rate": 1.1984754609016716e-05,
      "loss": 0.4651,
      "num_input_tokens_seen": 17633048,
      "step": 26975
    },
    {
      "epoch": 14.140461215932914,
      "grad_norm": 0.39414113759994507,
      "learning_rate": 1.1974993438602785e-05,
      "loss": 0.4909,
      "num_input_tokens_seen": 17636472,
      "step": 26980
    },
    {
      "epoch": 14.14308176100629,
      "grad_norm": 0.8795304298400879,
      "learning_rate": 1.1965234992871641e-05,
      "loss": 0.5441,
      "num_input_tokens_seen": 17640120,
      "step": 26985
    },
    {
      "epoch": 14.145702306079665,
      "grad_norm": 0.3620947003364563,
      "learning_rate": 1.195547927386462e-05,
      "loss": 0.553,
      "num_input_tokens_seen": 17643704,
      "step": 26990
    },
    {
      "epoch": 14.14832285115304,
      "grad_norm": 0.42849716544151306,
      "learning_rate": 1.1945726283622514e-05,
      "loss": 0.5421,
      "num_input_tokens_seen": 17646968,
      "step": 26995
    },
    {
      "epoch": 14.150943396226415,
      "grad_norm": 0.5936944484710693,
      "learning_rate": 1.1935976024185547e-05,
      "loss": 0.4395,
      "num_input_tokens_seen": 17650552,
      "step": 27000
    },
    {
      "epoch": 14.15356394129979,
      "grad_norm": 0.5433226823806763,
      "learning_rate": 1.1926228497593341e-05,
      "loss": 0.4282,
      "num_input_tokens_seen": 17653528,
      "step": 27005
    },
    {
      "epoch": 14.156184486373165,
      "grad_norm": 0.3185966908931732,
      "learning_rate": 1.1916483705884985e-05,
      "loss": 0.4009,
      "num_input_tokens_seen": 17656536,
      "step": 27010
    },
    {
      "epoch": 14.15880503144654,
      "grad_norm": 0.5793156623840332,
      "learning_rate": 1.1906741651098983e-05,
      "loss": 0.3859,
      "num_input_tokens_seen": 17662200,
      "step": 27015
    },
    {
      "epoch": 14.161425576519916,
      "grad_norm": 1.0796562433242798,
      "learning_rate": 1.1897002335273239e-05,
      "loss": 0.3849,
      "num_input_tokens_seen": 17665592,
      "step": 27020
    },
    {
      "epoch": 14.164046121593291,
      "grad_norm": 0.4660082161426544,
      "learning_rate": 1.1887265760445121e-05,
      "loss": 0.5074,
      "num_input_tokens_seen": 17668376,
      "step": 27025
    },
    {
      "epoch": 14.166666666666666,
      "grad_norm": 0.39520013332366943,
      "learning_rate": 1.1877531928651417e-05,
      "loss": 0.3794,
      "num_input_tokens_seen": 17672760,
      "step": 27030
    },
    {
      "epoch": 14.169287211740041,
      "grad_norm": 0.6397591829299927,
      "learning_rate": 1.1867800841928313e-05,
      "loss": 0.4431,
      "num_input_tokens_seen": 17675480,
      "step": 27035
    },
    {
      "epoch": 14.171907756813416,
      "grad_norm": 0.8135467767715454,
      "learning_rate": 1.1858072502311465e-05,
      "loss": 0.4424,
      "num_input_tokens_seen": 17678360,
      "step": 27040
    },
    {
      "epoch": 14.174528301886792,
      "grad_norm": 0.36501744389533997,
      "learning_rate": 1.1848346911835908e-05,
      "loss": 0.3466,
      "num_input_tokens_seen": 17682040,
      "step": 27045
    },
    {
      "epoch": 14.177148846960169,
      "grad_norm": 0.508715033531189,
      "learning_rate": 1.1838624072536137e-05,
      "loss": 0.5462,
      "num_input_tokens_seen": 17685432,
      "step": 27050
    },
    {
      "epoch": 14.179769392033544,
      "grad_norm": 0.40418577194213867,
      "learning_rate": 1.1828903986446055e-05,
      "loss": 0.4489,
      "num_input_tokens_seen": 17688728,
      "step": 27055
    },
    {
      "epoch": 14.182389937106919,
      "grad_norm": 0.7694272398948669,
      "learning_rate": 1.1819186655599007e-05,
      "loss": 0.4897,
      "num_input_tokens_seen": 17692216,
      "step": 27060
    },
    {
      "epoch": 14.185010482180294,
      "grad_norm": 0.3318481147289276,
      "learning_rate": 1.1809472082027734e-05,
      "loss": 0.3446,
      "num_input_tokens_seen": 17696184,
      "step": 27065
    },
    {
      "epoch": 14.18763102725367,
      "grad_norm": 0.5446181893348694,
      "learning_rate": 1.1799760267764407e-05,
      "loss": 0.3751,
      "num_input_tokens_seen": 17698808,
      "step": 27070
    },
    {
      "epoch": 14.190251572327044,
      "grad_norm": 0.7001630663871765,
      "learning_rate": 1.1790051214840634e-05,
      "loss": 0.5381,
      "num_input_tokens_seen": 17702648,
      "step": 27075
    },
    {
      "epoch": 14.19287211740042,
      "grad_norm": 0.41515812277793884,
      "learning_rate": 1.1780344925287443e-05,
      "loss": 0.3604,
      "num_input_tokens_seen": 17705464,
      "step": 27080
    },
    {
      "epoch": 14.195492662473795,
      "grad_norm": 0.4955054819583893,
      "learning_rate": 1.1770641401135282e-05,
      "loss": 0.5533,
      "num_input_tokens_seen": 17707864,
      "step": 27085
    },
    {
      "epoch": 14.19811320754717,
      "grad_norm": 0.47273123264312744,
      "learning_rate": 1.1760940644414003e-05,
      "loss": 0.5059,
      "num_input_tokens_seen": 17711736,
      "step": 27090
    },
    {
      "epoch": 14.200733752620545,
      "grad_norm": 0.4933779239654541,
      "learning_rate": 1.1751242657152914e-05,
      "loss": 0.3622,
      "num_input_tokens_seen": 17715992,
      "step": 27095
    },
    {
      "epoch": 14.20335429769392,
      "grad_norm": 0.5368130803108215,
      "learning_rate": 1.1741547441380698e-05,
      "loss": 0.2834,
      "num_input_tokens_seen": 17718616,
      "step": 27100
    },
    {
      "epoch": 14.205974842767295,
      "grad_norm": 0.7425287961959839,
      "learning_rate": 1.17318549991255e-05,
      "loss": 0.3651,
      "num_input_tokens_seen": 17721880,
      "step": 27105
    },
    {
      "epoch": 14.20859538784067,
      "grad_norm": 0.6448079347610474,
      "learning_rate": 1.1722165332414863e-05,
      "loss": 0.5398,
      "num_input_tokens_seen": 17724376,
      "step": 27110
    },
    {
      "epoch": 14.211215932914046,
      "grad_norm": 0.4988054037094116,
      "learning_rate": 1.1712478443275768e-05,
      "loss": 0.468,
      "num_input_tokens_seen": 17727448,
      "step": 27115
    },
    {
      "epoch": 14.213836477987421,
      "grad_norm": 0.4109390079975128,
      "learning_rate": 1.170279433373459e-05,
      "loss": 0.4169,
      "num_input_tokens_seen": 17730808,
      "step": 27120
    },
    {
      "epoch": 14.216457023060796,
      "grad_norm": 0.4298169016838074,
      "learning_rate": 1.1693113005817125e-05,
      "loss": 0.4455,
      "num_input_tokens_seen": 17733464,
      "step": 27125
    },
    {
      "epoch": 14.219077568134171,
      "grad_norm": 0.32600411772727966,
      "learning_rate": 1.1683434461548606e-05,
      "loss": 0.3227,
      "num_input_tokens_seen": 17736600,
      "step": 27130
    },
    {
      "epoch": 14.221698113207546,
      "grad_norm": 0.6025157570838928,
      "learning_rate": 1.1673758702953675e-05,
      "loss": 0.4279,
      "num_input_tokens_seen": 17739608,
      "step": 27135
    },
    {
      "epoch": 14.224318658280922,
      "grad_norm": 0.40271082520484924,
      "learning_rate": 1.1664085732056398e-05,
      "loss": 0.424,
      "num_input_tokens_seen": 17742840,
      "step": 27140
    },
    {
      "epoch": 14.226939203354299,
      "grad_norm": 1.117613434791565,
      "learning_rate": 1.1654415550880243e-05,
      "loss": 0.3539,
      "num_input_tokens_seen": 17744824,
      "step": 27145
    },
    {
      "epoch": 14.229559748427674,
      "grad_norm": 0.6860175132751465,
      "learning_rate": 1.1644748161448088e-05,
      "loss": 0.4597,
      "num_input_tokens_seen": 17747544,
      "step": 27150
    },
    {
      "epoch": 14.232180293501049,
      "grad_norm": 0.5138912200927734,
      "learning_rate": 1.1635083565782257e-05,
      "loss": 0.3494,
      "num_input_tokens_seen": 17750840,
      "step": 27155
    },
    {
      "epoch": 14.234800838574424,
      "grad_norm": 0.4855077266693115,
      "learning_rate": 1.1625421765904465e-05,
      "loss": 0.6027,
      "num_input_tokens_seen": 17753432,
      "step": 27160
    },
    {
      "epoch": 14.2374213836478,
      "grad_norm": 0.6921382546424866,
      "learning_rate": 1.1615762763835858e-05,
      "loss": 0.4484,
      "num_input_tokens_seen": 17756536,
      "step": 27165
    },
    {
      "epoch": 14.240041928721174,
      "grad_norm": 0.20938435196876526,
      "learning_rate": 1.1606106561596993e-05,
      "loss": 0.2959,
      "num_input_tokens_seen": 17759896,
      "step": 27170
    },
    {
      "epoch": 14.24266247379455,
      "grad_norm": 0.5892506241798401,
      "learning_rate": 1.1596453161207829e-05,
      "loss": 0.5257,
      "num_input_tokens_seen": 17763352,
      "step": 27175
    },
    {
      "epoch": 14.245283018867925,
      "grad_norm": 0.4826328158378601,
      "learning_rate": 1.1586802564687741e-05,
      "loss": 0.4244,
      "num_input_tokens_seen": 17765816,
      "step": 27180
    },
    {
      "epoch": 14.2479035639413,
      "grad_norm": 0.4740297496318817,
      "learning_rate": 1.1577154774055528e-05,
      "loss": 0.3683,
      "num_input_tokens_seen": 17768056,
      "step": 27185
    },
    {
      "epoch": 14.250524109014675,
      "grad_norm": 0.9306002855300903,
      "learning_rate": 1.1567509791329401e-05,
      "loss": 0.4044,
      "num_input_tokens_seen": 17771896,
      "step": 27190
    },
    {
      "epoch": 14.25314465408805,
      "grad_norm": 0.42934757471084595,
      "learning_rate": 1.155786761852699e-05,
      "loss": 0.3494,
      "num_input_tokens_seen": 17775096,
      "step": 27195
    },
    {
      "epoch": 14.255765199161425,
      "grad_norm": 0.4244072735309601,
      "learning_rate": 1.1548228257665317e-05,
      "loss": 0.4196,
      "num_input_tokens_seen": 17778392,
      "step": 27200
    },
    {
      "epoch": 14.2583857442348,
      "grad_norm": 0.513404130935669,
      "learning_rate": 1.1538591710760815e-05,
      "loss": 0.3906,
      "num_input_tokens_seen": 17781176,
      "step": 27205
    },
    {
      "epoch": 14.261006289308176,
      "grad_norm": 0.481317400932312,
      "learning_rate": 1.1528957979829355e-05,
      "loss": 0.6526,
      "num_input_tokens_seen": 17785272,
      "step": 27210
    },
    {
      "epoch": 14.26362683438155,
      "grad_norm": 0.39812397956848145,
      "learning_rate": 1.1519327066886198e-05,
      "loss": 0.4939,
      "num_input_tokens_seen": 17788280,
      "step": 27215
    },
    {
      "epoch": 14.266247379454926,
      "grad_norm": 0.4648493528366089,
      "learning_rate": 1.1509698973946032e-05,
      "loss": 0.4139,
      "num_input_tokens_seen": 17791064,
      "step": 27220
    },
    {
      "epoch": 14.268867924528301,
      "grad_norm": 0.34333789348602295,
      "learning_rate": 1.1500073703022926e-05,
      "loss": 0.2996,
      "num_input_tokens_seen": 17793656,
      "step": 27225
    },
    {
      "epoch": 14.271488469601676,
      "grad_norm": 0.47125235199928284,
      "learning_rate": 1.1490451256130394e-05,
      "loss": 0.4522,
      "num_input_tokens_seen": 17796536,
      "step": 27230
    },
    {
      "epoch": 14.274109014675052,
      "grad_norm": 0.39777499437332153,
      "learning_rate": 1.1480831635281328e-05,
      "loss": 0.4996,
      "num_input_tokens_seen": 17801304,
      "step": 27235
    },
    {
      "epoch": 14.276729559748428,
      "grad_norm": 0.4725640118122101,
      "learning_rate": 1.1471214842488046e-05,
      "loss": 0.4123,
      "num_input_tokens_seen": 17804472,
      "step": 27240
    },
    {
      "epoch": 14.279350104821804,
      "grad_norm": 0.5877180099487305,
      "learning_rate": 1.1461600879762277e-05,
      "loss": 0.3756,
      "num_input_tokens_seen": 17807928,
      "step": 27245
    },
    {
      "epoch": 14.281970649895179,
      "grad_norm": 1.034105896949768,
      "learning_rate": 1.1451989749115157e-05,
      "loss": 0.5647,
      "num_input_tokens_seen": 17810680,
      "step": 27250
    },
    {
      "epoch": 14.284591194968554,
      "grad_norm": 0.3358656167984009,
      "learning_rate": 1.1442381452557219e-05,
      "loss": 0.371,
      "num_input_tokens_seen": 17814008,
      "step": 27255
    },
    {
      "epoch": 14.28721174004193,
      "grad_norm": 0.6835769414901733,
      "learning_rate": 1.1432775992098397e-05,
      "loss": 0.4697,
      "num_input_tokens_seen": 17816440,
      "step": 27260
    },
    {
      "epoch": 14.289832285115304,
      "grad_norm": 0.45704516768455505,
      "learning_rate": 1.1423173369748056e-05,
      "loss": 0.3606,
      "num_input_tokens_seen": 17819256,
      "step": 27265
    },
    {
      "epoch": 14.29245283018868,
      "grad_norm": 0.4124678075313568,
      "learning_rate": 1.1413573587514952e-05,
      "loss": 0.3739,
      "num_input_tokens_seen": 17822232,
      "step": 27270
    },
    {
      "epoch": 14.295073375262055,
      "grad_norm": 0.5793915390968323,
      "learning_rate": 1.140397664740726e-05,
      "loss": 0.4617,
      "num_input_tokens_seen": 17825432,
      "step": 27275
    },
    {
      "epoch": 14.29769392033543,
      "grad_norm": 0.69263756275177,
      "learning_rate": 1.1394382551432534e-05,
      "loss": 0.5324,
      "num_input_tokens_seen": 17828952,
      "step": 27280
    },
    {
      "epoch": 14.300314465408805,
      "grad_norm": 0.5487104654312134,
      "learning_rate": 1.1384791301597767e-05,
      "loss": 0.4759,
      "num_input_tokens_seen": 17831672,
      "step": 27285
    },
    {
      "epoch": 14.30293501048218,
      "grad_norm": 0.3698110580444336,
      "learning_rate": 1.1375202899909317e-05,
      "loss": 0.4385,
      "num_input_tokens_seen": 17835384,
      "step": 27290
    },
    {
      "epoch": 14.305555555555555,
      "grad_norm": 0.4564119875431061,
      "learning_rate": 1.1365617348372984e-05,
      "loss": 0.5143,
      "num_input_tokens_seen": 17838552,
      "step": 27295
    },
    {
      "epoch": 14.30817610062893,
      "grad_norm": 0.7359466552734375,
      "learning_rate": 1.135603464899396e-05,
      "loss": 0.3883,
      "num_input_tokens_seen": 17841080,
      "step": 27300
    },
    {
      "epoch": 14.310796645702306,
      "grad_norm": 0.45534881949424744,
      "learning_rate": 1.1346454803776821e-05,
      "loss": 0.4195,
      "num_input_tokens_seen": 17847224,
      "step": 27305
    },
    {
      "epoch": 14.31341719077568,
      "grad_norm": 0.4833088517189026,
      "learning_rate": 1.1336877814725566e-05,
      "loss": 0.3695,
      "num_input_tokens_seen": 17850840,
      "step": 27310
    },
    {
      "epoch": 14.316037735849056,
      "grad_norm": 0.5437528491020203,
      "learning_rate": 1.1327303683843607e-05,
      "loss": 0.5003,
      "num_input_tokens_seen": 17853176,
      "step": 27315
    },
    {
      "epoch": 14.318658280922431,
      "grad_norm": 0.3403114676475525,
      "learning_rate": 1.1317732413133723e-05,
      "loss": 0.4012,
      "num_input_tokens_seen": 17856280,
      "step": 27320
    },
    {
      "epoch": 14.321278825995806,
      "grad_norm": 0.298173725605011,
      "learning_rate": 1.1308164004598118e-05,
      "loss": 0.3948,
      "num_input_tokens_seen": 17859064,
      "step": 27325
    },
    {
      "epoch": 14.323899371069182,
      "grad_norm": 0.4214475154876709,
      "learning_rate": 1.129859846023841e-05,
      "loss": 0.5184,
      "num_input_tokens_seen": 17861880,
      "step": 27330
    },
    {
      "epoch": 14.326519916142558,
      "grad_norm": 0.5101130604743958,
      "learning_rate": 1.1289035782055582e-05,
      "loss": 0.5621,
      "num_input_tokens_seen": 17868120,
      "step": 27335
    },
    {
      "epoch": 14.329140461215934,
      "grad_norm": 0.5966123938560486,
      "learning_rate": 1.1279475972050044e-05,
      "loss": 0.5887,
      "num_input_tokens_seen": 17871896,
      "step": 27340
    },
    {
      "epoch": 14.331761006289309,
      "grad_norm": 0.4028466045856476,
      "learning_rate": 1.1269919032221612e-05,
      "loss": 0.3904,
      "num_input_tokens_seen": 17875512,
      "step": 27345
    },
    {
      "epoch": 14.334381551362684,
      "grad_norm": 0.5723965167999268,
      "learning_rate": 1.1260364964569466e-05,
      "loss": 0.4875,
      "num_input_tokens_seen": 17879000,
      "step": 27350
    },
    {
      "epoch": 14.33700209643606,
      "grad_norm": 0.8280124664306641,
      "learning_rate": 1.125081377109223e-05,
      "loss": 0.5319,
      "num_input_tokens_seen": 17881368,
      "step": 27355
    },
    {
      "epoch": 14.339622641509434,
      "grad_norm": 0.4627526104450226,
      "learning_rate": 1.1241265453787882e-05,
      "loss": 0.5832,
      "num_input_tokens_seen": 17884408,
      "step": 27360
    },
    {
      "epoch": 14.34224318658281,
      "grad_norm": 0.7186830043792725,
      "learning_rate": 1.1231720014653837e-05,
      "loss": 0.6358,
      "num_input_tokens_seen": 17887224,
      "step": 27365
    },
    {
      "epoch": 14.344863731656185,
      "grad_norm": 0.6626065969467163,
      "learning_rate": 1.1222177455686896e-05,
      "loss": 0.3797,
      "num_input_tokens_seen": 17889720,
      "step": 27370
    },
    {
      "epoch": 14.34748427672956,
      "grad_norm": 0.2836547791957855,
      "learning_rate": 1.1212637778883237e-05,
      "loss": 0.346,
      "num_input_tokens_seen": 17892888,
      "step": 27375
    },
    {
      "epoch": 14.350104821802935,
      "grad_norm": 0.9128683805465698,
      "learning_rate": 1.1203100986238468e-05,
      "loss": 0.4032,
      "num_input_tokens_seen": 17896056,
      "step": 27380
    },
    {
      "epoch": 14.35272536687631,
      "grad_norm": 0.41511332988739014,
      "learning_rate": 1.1193567079747561e-05,
      "loss": 0.4141,
      "num_input_tokens_seen": 17898840,
      "step": 27385
    },
    {
      "epoch": 14.355345911949685,
      "grad_norm": 0.6002518534660339,
      "learning_rate": 1.1184036061404907e-05,
      "loss": 0.4723,
      "num_input_tokens_seen": 17901560,
      "step": 27390
    },
    {
      "epoch": 14.35796645702306,
      "grad_norm": 0.3524358570575714,
      "learning_rate": 1.117450793320429e-05,
      "loss": 0.4119,
      "num_input_tokens_seen": 17905432,
      "step": 27395
    },
    {
      "epoch": 14.360587002096436,
      "grad_norm": 0.4459671974182129,
      "learning_rate": 1.1164982697138893e-05,
      "loss": 0.4083,
      "num_input_tokens_seen": 17908088,
      "step": 27400
    },
    {
      "epoch": 14.36320754716981,
      "grad_norm": 0.6651037335395813,
      "learning_rate": 1.1155460355201267e-05,
      "loss": 0.4411,
      "num_input_tokens_seen": 17910616,
      "step": 27405
    },
    {
      "epoch": 14.365828092243186,
      "grad_norm": 0.6821874976158142,
      "learning_rate": 1.1145940909383399e-05,
      "loss": 0.3768,
      "num_input_tokens_seen": 17913400,
      "step": 27410
    },
    {
      "epoch": 14.368448637316561,
      "grad_norm": 0.3221481442451477,
      "learning_rate": 1.1136424361676626e-05,
      "loss": 0.4794,
      "num_input_tokens_seen": 17916600,
      "step": 27415
    },
    {
      "epoch": 14.371069182389936,
      "grad_norm": 1.0152870416641235,
      "learning_rate": 1.1126910714071711e-05,
      "loss": 0.4297,
      "num_input_tokens_seen": 17919160,
      "step": 27420
    },
    {
      "epoch": 14.373689727463312,
      "grad_norm": 0.4983786344528198,
      "learning_rate": 1.1117399968558806e-05,
      "loss": 0.3689,
      "num_input_tokens_seen": 17921912,
      "step": 27425
    },
    {
      "epoch": 14.376310272536688,
      "grad_norm": 0.41473203897476196,
      "learning_rate": 1.1107892127127454e-05,
      "loss": 0.4832,
      "num_input_tokens_seen": 17926040,
      "step": 27430
    },
    {
      "epoch": 14.378930817610064,
      "grad_norm": 0.2876443862915039,
      "learning_rate": 1.109838719176658e-05,
      "loss": 0.4285,
      "num_input_tokens_seen": 17929656,
      "step": 27435
    },
    {
      "epoch": 14.381551362683439,
      "grad_norm": 0.36643290519714355,
      "learning_rate": 1.1088885164464497e-05,
      "loss": 0.3626,
      "num_input_tokens_seen": 17935704,
      "step": 27440
    },
    {
      "epoch": 14.384171907756814,
      "grad_norm": 0.6443442106246948,
      "learning_rate": 1.1079386047208932e-05,
      "loss": 0.5024,
      "num_input_tokens_seen": 17938328,
      "step": 27445
    },
    {
      "epoch": 14.38679245283019,
      "grad_norm": 0.4948248565196991,
      "learning_rate": 1.1069889841986991e-05,
      "loss": 0.4004,
      "num_input_tokens_seen": 17941240,
      "step": 27450
    },
    {
      "epoch": 14.389412997903564,
      "grad_norm": 0.3772377371788025,
      "learning_rate": 1.1060396550785182e-05,
      "loss": 0.5296,
      "num_input_tokens_seen": 17944728,
      "step": 27455
    },
    {
      "epoch": 14.39203354297694,
      "grad_norm": 0.5993074774742126,
      "learning_rate": 1.1050906175589382e-05,
      "loss": 0.478,
      "num_input_tokens_seen": 17948536,
      "step": 27460
    },
    {
      "epoch": 14.394654088050315,
      "grad_norm": 0.509611964225769,
      "learning_rate": 1.1041418718384863e-05,
      "loss": 0.5495,
      "num_input_tokens_seen": 17951448,
      "step": 27465
    },
    {
      "epoch": 14.39727463312369,
      "grad_norm": 0.5077551603317261,
      "learning_rate": 1.1031934181156298e-05,
      "loss": 0.4702,
      "num_input_tokens_seen": 17954776,
      "step": 27470
    },
    {
      "epoch": 14.399895178197065,
      "grad_norm": 0.5191969871520996,
      "learning_rate": 1.1022452565887747e-05,
      "loss": 0.3521,
      "num_input_tokens_seen": 17958712,
      "step": 27475
    },
    {
      "epoch": 14.40251572327044,
      "grad_norm": 0.349624365568161,
      "learning_rate": 1.1012973874562657e-05,
      "loss": 0.3538,
      "num_input_tokens_seen": 17961464,
      "step": 27480
    },
    {
      "epoch": 14.405136268343815,
      "grad_norm": 0.48243558406829834,
      "learning_rate": 1.1003498109163865e-05,
      "loss": 0.4294,
      "num_input_tokens_seen": 17965176,
      "step": 27485
    },
    {
      "epoch": 14.40775681341719,
      "grad_norm": 0.7552982568740845,
      "learning_rate": 1.0994025271673592e-05,
      "loss": 0.4991,
      "num_input_tokens_seen": 17968440,
      "step": 27490
    },
    {
      "epoch": 14.410377358490566,
      "grad_norm": 0.9021921157836914,
      "learning_rate": 1.0984555364073431e-05,
      "loss": 0.448,
      "num_input_tokens_seen": 17971416,
      "step": 27495
    },
    {
      "epoch": 14.41299790356394,
      "grad_norm": 0.6690434813499451,
      "learning_rate": 1.0975088388344393e-05,
      "loss": 0.5601,
      "num_input_tokens_seen": 17974360,
      "step": 27500
    },
    {
      "epoch": 14.415618448637316,
      "grad_norm": 0.28054162859916687,
      "learning_rate": 1.096562434646686e-05,
      "loss": 0.4209,
      "num_input_tokens_seen": 17977912,
      "step": 27505
    },
    {
      "epoch": 14.418238993710691,
      "grad_norm": 0.34423187375068665,
      "learning_rate": 1.0956163240420611e-05,
      "loss": 0.3299,
      "num_input_tokens_seen": 17980792,
      "step": 27510
    },
    {
      "epoch": 14.420859538784066,
      "grad_norm": 1.189975619316101,
      "learning_rate": 1.094670507218479e-05,
      "loss": 0.3977,
      "num_input_tokens_seen": 17983416,
      "step": 27515
    },
    {
      "epoch": 14.423480083857442,
      "grad_norm": 0.5717096328735352,
      "learning_rate": 1.093724984373793e-05,
      "loss": 0.4252,
      "num_input_tokens_seen": 17986328,
      "step": 27520
    },
    {
      "epoch": 14.426100628930818,
      "grad_norm": 0.5626137852668762,
      "learning_rate": 1.0927797557057968e-05,
      "loss": 0.4002,
      "num_input_tokens_seen": 17990584,
      "step": 27525
    },
    {
      "epoch": 14.428721174004194,
      "grad_norm": 0.33060890436172485,
      "learning_rate": 1.0918348214122209e-05,
      "loss": 0.4249,
      "num_input_tokens_seen": 17993848,
      "step": 27530
    },
    {
      "epoch": 14.431341719077569,
      "grad_norm": 0.4861377775669098,
      "learning_rate": 1.0908901816907366e-05,
      "loss": 0.4564,
      "num_input_tokens_seen": 17997208,
      "step": 27535
    },
    {
      "epoch": 14.433962264150944,
      "grad_norm": 0.8819321990013123,
      "learning_rate": 1.08994583673895e-05,
      "loss": 0.5036,
      "num_input_tokens_seen": 18000056,
      "step": 27540
    },
    {
      "epoch": 14.43658280922432,
      "grad_norm": 0.425143301486969,
      "learning_rate": 1.0890017867544068e-05,
      "loss": 0.5861,
      "num_input_tokens_seen": 18002968,
      "step": 27545
    },
    {
      "epoch": 14.439203354297694,
      "grad_norm": 0.3401429057121277,
      "learning_rate": 1.088058031934592e-05,
      "loss": 0.3893,
      "num_input_tokens_seen": 18006104,
      "step": 27550
    },
    {
      "epoch": 14.44182389937107,
      "grad_norm": 0.45636528730392456,
      "learning_rate": 1.087114572476929e-05,
      "loss": 0.4378,
      "num_input_tokens_seen": 18010040,
      "step": 27555
    },
    {
      "epoch": 14.444444444444445,
      "grad_norm": 0.3992796540260315,
      "learning_rate": 1.0861714085787783e-05,
      "loss": 0.3776,
      "num_input_tokens_seen": 18013720,
      "step": 27560
    },
    {
      "epoch": 14.44706498951782,
      "grad_norm": 0.36175018548965454,
      "learning_rate": 1.0852285404374402e-05,
      "loss": 0.3326,
      "num_input_tokens_seen": 18016696,
      "step": 27565
    },
    {
      "epoch": 14.449685534591195,
      "grad_norm": 0.29526397585868835,
      "learning_rate": 1.0842859682501505e-05,
      "loss": 0.5399,
      "num_input_tokens_seen": 18020024,
      "step": 27570
    },
    {
      "epoch": 14.45230607966457,
      "grad_norm": 0.8707574605941772,
      "learning_rate": 1.0833436922140841e-05,
      "loss": 0.3655,
      "num_input_tokens_seen": 18023352,
      "step": 27575
    },
    {
      "epoch": 14.454926624737945,
      "grad_norm": 0.5863589644432068,
      "learning_rate": 1.0824017125263555e-05,
      "loss": 0.4418,
      "num_input_tokens_seen": 18026648,
      "step": 27580
    },
    {
      "epoch": 14.45754716981132,
      "grad_norm": 0.42198577523231506,
      "learning_rate": 1.0814600293840154e-05,
      "loss": 0.3586,
      "num_input_tokens_seen": 18029624,
      "step": 27585
    },
    {
      "epoch": 14.460167714884696,
      "grad_norm": 0.4460310935974121,
      "learning_rate": 1.080518642984055e-05,
      "loss": 0.5619,
      "num_input_tokens_seen": 18032376,
      "step": 27590
    },
    {
      "epoch": 14.46278825995807,
      "grad_norm": 0.4141225814819336,
      "learning_rate": 1.0795775535233988e-05,
      "loss": 0.4087,
      "num_input_tokens_seen": 18035448,
      "step": 27595
    },
    {
      "epoch": 14.465408805031446,
      "grad_norm": 0.40772148966789246,
      "learning_rate": 1.0786367611989145e-05,
      "loss": 0.4782,
      "num_input_tokens_seen": 18039576,
      "step": 27600
    },
    {
      "epoch": 14.468029350104821,
      "grad_norm": 0.6473666429519653,
      "learning_rate": 1.0776962662074031e-05,
      "loss": 0.3451,
      "num_input_tokens_seen": 18044952,
      "step": 27605
    },
    {
      "epoch": 14.470649895178196,
      "grad_norm": 0.506306529045105,
      "learning_rate": 1.0767560687456062e-05,
      "loss": 0.4369,
      "num_input_tokens_seen": 18048472,
      "step": 27610
    },
    {
      "epoch": 14.473270440251572,
      "grad_norm": 0.3165307641029358,
      "learning_rate": 1.0758161690102036e-05,
      "loss": 0.4387,
      "num_input_tokens_seen": 18052088,
      "step": 27615
    },
    {
      "epoch": 14.475890985324948,
      "grad_norm": 0.390256404876709,
      "learning_rate": 1.0748765671978095e-05,
      "loss": 0.4472,
      "num_input_tokens_seen": 18055000,
      "step": 27620
    },
    {
      "epoch": 14.478511530398324,
      "grad_norm": 0.35528674721717834,
      "learning_rate": 1.0739372635049797e-05,
      "loss": 0.4135,
      "num_input_tokens_seen": 18058712,
      "step": 27625
    },
    {
      "epoch": 14.481132075471699,
      "grad_norm": 0.895110547542572,
      "learning_rate": 1.0729982581282042e-05,
      "loss": 0.3221,
      "num_input_tokens_seen": 18062616,
      "step": 27630
    },
    {
      "epoch": 14.483752620545074,
      "grad_norm": 0.7162454128265381,
      "learning_rate": 1.0720595512639129e-05,
      "loss": 0.396,
      "num_input_tokens_seen": 18066360,
      "step": 27635
    },
    {
      "epoch": 14.48637316561845,
      "grad_norm": 0.6396105885505676,
      "learning_rate": 1.0711211431084728e-05,
      "loss": 0.4477,
      "num_input_tokens_seen": 18069848,
      "step": 27640
    },
    {
      "epoch": 14.488993710691824,
      "grad_norm": 0.3513193428516388,
      "learning_rate": 1.0701830338581887e-05,
      "loss": 0.428,
      "num_input_tokens_seen": 18072504,
      "step": 27645
    },
    {
      "epoch": 14.4916142557652,
      "grad_norm": 0.4553193747997284,
      "learning_rate": 1.069245223709301e-05,
      "loss": 0.3568,
      "num_input_tokens_seen": 18075960,
      "step": 27650
    },
    {
      "epoch": 14.494234800838575,
      "grad_norm": 0.33897581696510315,
      "learning_rate": 1.0683077128579904e-05,
      "loss": 0.4216,
      "num_input_tokens_seen": 18079608,
      "step": 27655
    },
    {
      "epoch": 14.49685534591195,
      "grad_norm": 0.39137759804725647,
      "learning_rate": 1.0673705015003718e-05,
      "loss": 0.555,
      "num_input_tokens_seen": 18082232,
      "step": 27660
    },
    {
      "epoch": 14.499475890985325,
      "grad_norm": 0.4025103747844696,
      "learning_rate": 1.0664335898325002e-05,
      "loss": 0.364,
      "num_input_tokens_seen": 18085432,
      "step": 27665
    },
    {
      "epoch": 14.5020964360587,
      "grad_norm": 0.4019279479980469,
      "learning_rate": 1.065496978050367e-05,
      "loss": 0.3815,
      "num_input_tokens_seen": 18089176,
      "step": 27670
    },
    {
      "epoch": 14.504716981132075,
      "grad_norm": 0.7143831849098206,
      "learning_rate": 1.0645606663499e-05,
      "loss": 0.515,
      "num_input_tokens_seen": 18092600,
      "step": 27675
    },
    {
      "epoch": 14.50733752620545,
      "grad_norm": 0.5080786347389221,
      "learning_rate": 1.0636246549269649e-05,
      "loss": 0.6035,
      "num_input_tokens_seen": 18096056,
      "step": 27680
    },
    {
      "epoch": 14.509958071278826,
      "grad_norm": 0.3422786593437195,
      "learning_rate": 1.062688943977366e-05,
      "loss": 0.5044,
      "num_input_tokens_seen": 18099192,
      "step": 27685
    },
    {
      "epoch": 14.5125786163522,
      "grad_norm": 0.30178025364875793,
      "learning_rate": 1.0617535336968413e-05,
      "loss": 0.4572,
      "num_input_tokens_seen": 18106744,
      "step": 27690
    },
    {
      "epoch": 14.515199161425576,
      "grad_norm": 0.48585042357444763,
      "learning_rate": 1.0608184242810698e-05,
      "loss": 0.4285,
      "num_input_tokens_seen": 18109592,
      "step": 27695
    },
    {
      "epoch": 14.517819706498951,
      "grad_norm": 0.6806744933128357,
      "learning_rate": 1.0598836159256642e-05,
      "loss": 0.4465,
      "num_input_tokens_seen": 18112632,
      "step": 27700
    },
    {
      "epoch": 14.520440251572326,
      "grad_norm": 0.4672451615333557,
      "learning_rate": 1.0589491088261763e-05,
      "loss": 0.4783,
      "num_input_tokens_seen": 18116824,
      "step": 27705
    },
    {
      "epoch": 14.523060796645701,
      "grad_norm": 0.6275388598442078,
      "learning_rate": 1.0580149031780945e-05,
      "loss": 0.5675,
      "num_input_tokens_seen": 18120504,
      "step": 27710
    },
    {
      "epoch": 14.525681341719078,
      "grad_norm": 1.0567694902420044,
      "learning_rate": 1.0570809991768447e-05,
      "loss": 0.3818,
      "num_input_tokens_seen": 18123768,
      "step": 27715
    },
    {
      "epoch": 14.528301886792454,
      "grad_norm": 0.6202723383903503,
      "learning_rate": 1.0561473970177873e-05,
      "loss": 0.3723,
      "num_input_tokens_seen": 18126584,
      "step": 27720
    },
    {
      "epoch": 14.530922431865829,
      "grad_norm": 0.4312213361263275,
      "learning_rate": 1.0552140968962232e-05,
      "loss": 0.3957,
      "num_input_tokens_seen": 18131480,
      "step": 27725
    },
    {
      "epoch": 14.533542976939204,
      "grad_norm": 0.4928460419178009,
      "learning_rate": 1.054281099007386e-05,
      "loss": 0.5591,
      "num_input_tokens_seen": 18135224,
      "step": 27730
    },
    {
      "epoch": 14.536163522012579,
      "grad_norm": 0.4125804901123047,
      "learning_rate": 1.0533484035464492e-05,
      "loss": 0.3424,
      "num_input_tokens_seen": 18138232,
      "step": 27735
    },
    {
      "epoch": 14.538784067085954,
      "grad_norm": 0.4520704448223114,
      "learning_rate": 1.0524160107085231e-05,
      "loss": 0.5144,
      "num_input_tokens_seen": 18142040,
      "step": 27740
    },
    {
      "epoch": 14.54140461215933,
      "grad_norm": 0.4949919283390045,
      "learning_rate": 1.0514839206886518e-05,
      "loss": 0.4365,
      "num_input_tokens_seen": 18144888,
      "step": 27745
    },
    {
      "epoch": 14.544025157232705,
      "grad_norm": 0.6226474046707153,
      "learning_rate": 1.0505521336818195e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 18147672,
      "step": 27750
    },
    {
      "epoch": 14.54664570230608,
      "grad_norm": 0.6386710405349731,
      "learning_rate": 1.0496206498829438e-05,
      "loss": 0.283,
      "num_input_tokens_seen": 18150808,
      "step": 27755
    },
    {
      "epoch": 14.549266247379455,
      "grad_norm": 0.37055855989456177,
      "learning_rate": 1.0486894694868814e-05,
      "loss": 0.4072,
      "num_input_tokens_seen": 18153400,
      "step": 27760
    },
    {
      "epoch": 14.55188679245283,
      "grad_norm": 0.32843419909477234,
      "learning_rate": 1.0477585926884246e-05,
      "loss": 0.3534,
      "num_input_tokens_seen": 18157432,
      "step": 27765
    },
    {
      "epoch": 14.554507337526205,
      "grad_norm": 0.9556228518486023,
      "learning_rate": 1.046828019682303e-05,
      "loss": 0.5575,
      "num_input_tokens_seen": 18160472,
      "step": 27770
    },
    {
      "epoch": 14.55712788259958,
      "grad_norm": 0.4511973559856415,
      "learning_rate": 1.0458977506631808e-05,
      "loss": 0.4303,
      "num_input_tokens_seen": 18163608,
      "step": 27775
    },
    {
      "epoch": 14.559748427672956,
      "grad_norm": 0.43717265129089355,
      "learning_rate": 1.0449677858256593e-05,
      "loss": 0.4038,
      "num_input_tokens_seen": 18167736,
      "step": 27780
    },
    {
      "epoch": 14.56236897274633,
      "grad_norm": 0.7349019050598145,
      "learning_rate": 1.0440381253642773e-05,
      "loss": 0.3664,
      "num_input_tokens_seen": 18171160,
      "step": 27785
    },
    {
      "epoch": 14.564989517819706,
      "grad_norm": 0.45869046449661255,
      "learning_rate": 1.0431087694735087e-05,
      "loss": 0.5581,
      "num_input_tokens_seen": 18174168,
      "step": 27790
    },
    {
      "epoch": 14.567610062893081,
      "grad_norm": 0.26612576842308044,
      "learning_rate": 1.0421797183477644e-05,
      "loss": 0.5681,
      "num_input_tokens_seen": 18177080,
      "step": 27795
    },
    {
      "epoch": 14.570230607966456,
      "grad_norm": 0.29253196716308594,
      "learning_rate": 1.041250972181392e-05,
      "loss": 0.455,
      "num_input_tokens_seen": 18181976,
      "step": 27800
    },
    {
      "epoch": 14.572851153039831,
      "grad_norm": 0.3852434754371643,
      "learning_rate": 1.0403225311686743e-05,
      "loss": 0.4877,
      "num_input_tokens_seen": 18186168,
      "step": 27805
    },
    {
      "epoch": 14.575471698113208,
      "grad_norm": 0.4867000877857208,
      "learning_rate": 1.0393943955038293e-05,
      "loss": 0.3788,
      "num_input_tokens_seen": 18189656,
      "step": 27810
    },
    {
      "epoch": 14.578092243186584,
      "grad_norm": 0.519433856010437,
      "learning_rate": 1.038466565381013e-05,
      "loss": 0.3363,
      "num_input_tokens_seen": 18191768,
      "step": 27815
    },
    {
      "epoch": 14.580712788259959,
      "grad_norm": 0.2730281352996826,
      "learning_rate": 1.0375390409943175e-05,
      "loss": 0.3752,
      "num_input_tokens_seen": 18194296,
      "step": 27820
    },
    {
      "epoch": 14.583333333333334,
      "grad_norm": 0.5267091393470764,
      "learning_rate": 1.0366118225377708e-05,
      "loss": 0.4753,
      "num_input_tokens_seen": 18197528,
      "step": 27825
    },
    {
      "epoch": 14.585953878406709,
      "grad_norm": 0.4490439295768738,
      "learning_rate": 1.0356849102053353e-05,
      "loss": 0.3246,
      "num_input_tokens_seen": 18200600,
      "step": 27830
    },
    {
      "epoch": 14.588574423480084,
      "grad_norm": 0.6496240496635437,
      "learning_rate": 1.0347583041909104e-05,
      "loss": 0.4501,
      "num_input_tokens_seen": 18203864,
      "step": 27835
    },
    {
      "epoch": 14.59119496855346,
      "grad_norm": 0.4179026782512665,
      "learning_rate": 1.0338320046883315e-05,
      "loss": 0.4605,
      "num_input_tokens_seen": 18207192,
      "step": 27840
    },
    {
      "epoch": 14.593815513626835,
      "grad_norm": 0.3262477219104767,
      "learning_rate": 1.0329060118913705e-05,
      "loss": 0.4311,
      "num_input_tokens_seen": 18211352,
      "step": 27845
    },
    {
      "epoch": 14.59643605870021,
      "grad_norm": 0.6960515379905701,
      "learning_rate": 1.0319803259937353e-05,
      "loss": 0.4236,
      "num_input_tokens_seen": 18214200,
      "step": 27850
    },
    {
      "epoch": 14.599056603773585,
      "grad_norm": 0.6705958843231201,
      "learning_rate": 1.0310549471890676e-05,
      "loss": 0.3997,
      "num_input_tokens_seen": 18216920,
      "step": 27855
    },
    {
      "epoch": 14.60167714884696,
      "grad_norm": 1.0218942165374756,
      "learning_rate": 1.0301298756709454e-05,
      "loss": 0.5558,
      "num_input_tokens_seen": 18220184,
      "step": 27860
    },
    {
      "epoch": 14.604297693920335,
      "grad_norm": 0.3305300176143646,
      "learning_rate": 1.0292051116328844e-05,
      "loss": 0.3497,
      "num_input_tokens_seen": 18224280,
      "step": 27865
    },
    {
      "epoch": 14.60691823899371,
      "grad_norm": 0.5957716703414917,
      "learning_rate": 1.028280655268334e-05,
      "loss": 0.3373,
      "num_input_tokens_seen": 18226744,
      "step": 27870
    },
    {
      "epoch": 14.609538784067086,
      "grad_norm": 0.5253322720527649,
      "learning_rate": 1.0273565067706806e-05,
      "loss": 0.5071,
      "num_input_tokens_seen": 18230232,
      "step": 27875
    },
    {
      "epoch": 14.61215932914046,
      "grad_norm": 0.515670895576477,
      "learning_rate": 1.0264326663332457e-05,
      "loss": 0.4289,
      "num_input_tokens_seen": 18234040,
      "step": 27880
    },
    {
      "epoch": 14.614779874213836,
      "grad_norm": 0.47343775629997253,
      "learning_rate": 1.0255091341492859e-05,
      "loss": 0.2013,
      "num_input_tokens_seen": 18236632,
      "step": 27885
    },
    {
      "epoch": 14.617400419287211,
      "grad_norm": 0.48668569326400757,
      "learning_rate": 1.0245859104119923e-05,
      "loss": 0.4034,
      "num_input_tokens_seen": 18239640,
      "step": 27890
    },
    {
      "epoch": 14.620020964360586,
      "grad_norm": 0.3799781799316406,
      "learning_rate": 1.0236629953144938e-05,
      "loss": 0.4333,
      "num_input_tokens_seen": 18242648,
      "step": 27895
    },
    {
      "epoch": 14.622641509433961,
      "grad_norm": 0.42967626452445984,
      "learning_rate": 1.0227403890498537e-05,
      "loss": 0.3995,
      "num_input_tokens_seen": 18246584,
      "step": 27900
    },
    {
      "epoch": 14.625262054507338,
      "grad_norm": 0.7654616832733154,
      "learning_rate": 1.021818091811072e-05,
      "loss": 0.5571,
      "num_input_tokens_seen": 18249400,
      "step": 27905
    },
    {
      "epoch": 14.627882599580714,
      "grad_norm": 0.4354036748409271,
      "learning_rate": 1.0208961037910813e-05,
      "loss": 0.5563,
      "num_input_tokens_seen": 18252952,
      "step": 27910
    },
    {
      "epoch": 14.630503144654089,
      "grad_norm": 0.42519935965538025,
      "learning_rate": 1.0199744251827503e-05,
      "loss": 0.5133,
      "num_input_tokens_seen": 18257176,
      "step": 27915
    },
    {
      "epoch": 14.633123689727464,
      "grad_norm": 0.36927810311317444,
      "learning_rate": 1.0190530561788845e-05,
      "loss": 0.4376,
      "num_input_tokens_seen": 18260280,
      "step": 27920
    },
    {
      "epoch": 14.635744234800839,
      "grad_norm": 0.44058772921562195,
      "learning_rate": 1.0181319969722236e-05,
      "loss": 0.4459,
      "num_input_tokens_seen": 18263480,
      "step": 27925
    },
    {
      "epoch": 14.638364779874214,
      "grad_norm": 0.5590951442718506,
      "learning_rate": 1.0172112477554441e-05,
      "loss": 0.3201,
      "num_input_tokens_seen": 18267128,
      "step": 27930
    },
    {
      "epoch": 14.64098532494759,
      "grad_norm": 0.29184871912002563,
      "learning_rate": 1.016290808721154e-05,
      "loss": 0.5322,
      "num_input_tokens_seen": 18270968,
      "step": 27935
    },
    {
      "epoch": 14.643605870020965,
      "grad_norm": 0.46007436513900757,
      "learning_rate": 1.0153706800619007e-05,
      "loss": 0.4887,
      "num_input_tokens_seen": 18273720,
      "step": 27940
    },
    {
      "epoch": 14.64622641509434,
      "grad_norm": 0.44005322456359863,
      "learning_rate": 1.0144508619701623e-05,
      "loss": 0.4232,
      "num_input_tokens_seen": 18276920,
      "step": 27945
    },
    {
      "epoch": 14.648846960167715,
      "grad_norm": 0.3911457657814026,
      "learning_rate": 1.013531354638356e-05,
      "loss": 0.5652,
      "num_input_tokens_seen": 18280952,
      "step": 27950
    },
    {
      "epoch": 14.65146750524109,
      "grad_norm": 0.4590955674648285,
      "learning_rate": 1.0126121582588316e-05,
      "loss": 0.6043,
      "num_input_tokens_seen": 18284120,
      "step": 27955
    },
    {
      "epoch": 14.654088050314465,
      "grad_norm": 0.38176101446151733,
      "learning_rate": 1.0116932730238754e-05,
      "loss": 0.4907,
      "num_input_tokens_seen": 18288120,
      "step": 27960
    },
    {
      "epoch": 14.65670859538784,
      "grad_norm": 0.47339966893196106,
      "learning_rate": 1.0107746991257063e-05,
      "loss": 0.4604,
      "num_input_tokens_seen": 18291224,
      "step": 27965
    },
    {
      "epoch": 14.659329140461216,
      "grad_norm": 0.25273561477661133,
      "learning_rate": 1.0098564367564811e-05,
      "loss": 0.3622,
      "num_input_tokens_seen": 18294520,
      "step": 27970
    },
    {
      "epoch": 14.66194968553459,
      "grad_norm": 0.42006638646125793,
      "learning_rate": 1.0089384861082876e-05,
      "loss": 0.3617,
      "num_input_tokens_seen": 18298296,
      "step": 27975
    },
    {
      "epoch": 14.664570230607966,
      "grad_norm": 0.5099436044692993,
      "learning_rate": 1.0080208473731523e-05,
      "loss": 0.4346,
      "num_input_tokens_seen": 18301016,
      "step": 27980
    },
    {
      "epoch": 14.667190775681341,
      "grad_norm": 0.4203747510910034,
      "learning_rate": 1.0071035207430352e-05,
      "loss": 0.4556,
      "num_input_tokens_seen": 18304728,
      "step": 27985
    },
    {
      "epoch": 14.669811320754716,
      "grad_norm": 0.6108990907669067,
      "learning_rate": 1.0061865064098285e-05,
      "loss": 0.4901,
      "num_input_tokens_seen": 18307736,
      "step": 27990
    },
    {
      "epoch": 14.672431865828091,
      "grad_norm": 0.3847079277038574,
      "learning_rate": 1.0052698045653625e-05,
      "loss": 0.3725,
      "num_input_tokens_seen": 18310744,
      "step": 27995
    },
    {
      "epoch": 14.675052410901468,
      "grad_norm": 1.141183853149414,
      "learning_rate": 1.0043534154014015e-05,
      "loss": 0.5257,
      "num_input_tokens_seen": 18313336,
      "step": 28000
    },
    {
      "epoch": 14.677672955974844,
      "grad_norm": 0.9705966114997864,
      "learning_rate": 1.003437339109642e-05,
      "loss": 0.5552,
      "num_input_tokens_seen": 18316120,
      "step": 28005
    },
    {
      "epoch": 14.680293501048219,
      "grad_norm": 0.2814468741416931,
      "learning_rate": 1.0025215758817183e-05,
      "loss": 0.4764,
      "num_input_tokens_seen": 18319480,
      "step": 28010
    },
    {
      "epoch": 14.682914046121594,
      "grad_norm": 0.40737155079841614,
      "learning_rate": 1.0016061259091958e-05,
      "loss": 0.321,
      "num_input_tokens_seen": 18323288,
      "step": 28015
    },
    {
      "epoch": 14.685534591194969,
      "grad_norm": 0.43980950117111206,
      "learning_rate": 1.000690989383577e-05,
      "loss": 0.3968,
      "num_input_tokens_seen": 18325752,
      "step": 28020
    },
    {
      "epoch": 14.688155136268344,
      "grad_norm": 0.46589162945747375,
      "learning_rate": 9.997761664962993e-06,
      "loss": 0.3696,
      "num_input_tokens_seen": 18330424,
      "step": 28025
    },
    {
      "epoch": 14.69077568134172,
      "grad_norm": 0.7189913392066956,
      "learning_rate": 9.988616574387314e-06,
      "loss": 0.4255,
      "num_input_tokens_seen": 18333336,
      "step": 28030
    },
    {
      "epoch": 14.693396226415095,
      "grad_norm": 0.3478826880455017,
      "learning_rate": 9.979474624021787e-06,
      "loss": 0.3896,
      "num_input_tokens_seen": 18336760,
      "step": 28035
    },
    {
      "epoch": 14.69601677148847,
      "grad_norm": 0.8756061792373657,
      "learning_rate": 9.970335815778815e-06,
      "loss": 0.4148,
      "num_input_tokens_seen": 18339288,
      "step": 28040
    },
    {
      "epoch": 14.698637316561845,
      "grad_norm": 0.32800373435020447,
      "learning_rate": 9.96120015157011e-06,
      "loss": 0.5806,
      "num_input_tokens_seen": 18343064,
      "step": 28045
    },
    {
      "epoch": 14.70125786163522,
      "grad_norm": 0.34726449847221375,
      "learning_rate": 9.952067633306767e-06,
      "loss": 0.4505,
      "num_input_tokens_seen": 18346584,
      "step": 28050
    },
    {
      "epoch": 14.703878406708595,
      "grad_norm": 0.4734272360801697,
      "learning_rate": 9.942938262899204e-06,
      "loss": 0.3719,
      "num_input_tokens_seen": 18349752,
      "step": 28055
    },
    {
      "epoch": 14.70649895178197,
      "grad_norm": 0.6985578536987305,
      "learning_rate": 9.93381204225717e-06,
      "loss": 0.3569,
      "num_input_tokens_seen": 18352632,
      "step": 28060
    },
    {
      "epoch": 14.709119496855346,
      "grad_norm": 0.6976146101951599,
      "learning_rate": 9.924688973289782e-06,
      "loss": 0.4436,
      "num_input_tokens_seen": 18356280,
      "step": 28065
    },
    {
      "epoch": 14.71174004192872,
      "grad_norm": 0.42573991417884827,
      "learning_rate": 9.915569057905463e-06,
      "loss": 0.3521,
      "num_input_tokens_seen": 18359512,
      "step": 28070
    },
    {
      "epoch": 14.714360587002096,
      "grad_norm": 0.795005202293396,
      "learning_rate": 9.906452298012003e-06,
      "loss": 0.4935,
      "num_input_tokens_seen": 18362072,
      "step": 28075
    },
    {
      "epoch": 14.716981132075471,
      "grad_norm": 0.6265943646430969,
      "learning_rate": 9.897338695516529e-06,
      "loss": 0.4545,
      "num_input_tokens_seen": 18365400,
      "step": 28080
    },
    {
      "epoch": 14.719601677148846,
      "grad_norm": 0.26583367586135864,
      "learning_rate": 9.888228252325507e-06,
      "loss": 0.4644,
      "num_input_tokens_seen": 18369176,
      "step": 28085
    },
    {
      "epoch": 14.722222222222221,
      "grad_norm": 0.6658518314361572,
      "learning_rate": 9.87912097034473e-06,
      "loss": 0.3572,
      "num_input_tokens_seen": 18372472,
      "step": 28090
    },
    {
      "epoch": 14.724842767295598,
      "grad_norm": 0.4397289454936981,
      "learning_rate": 9.870016851479327e-06,
      "loss": 0.4567,
      "num_input_tokens_seen": 18375608,
      "step": 28095
    },
    {
      "epoch": 14.727463312368974,
      "grad_norm": 0.5431718230247498,
      "learning_rate": 9.860915897633791e-06,
      "loss": 0.5382,
      "num_input_tokens_seen": 18378424,
      "step": 28100
    },
    {
      "epoch": 14.730083857442349,
      "grad_norm": 0.4771803915500641,
      "learning_rate": 9.85181811071193e-06,
      "loss": 0.5494,
      "num_input_tokens_seen": 18383544,
      "step": 28105
    },
    {
      "epoch": 14.732704402515724,
      "grad_norm": 0.4819604456424713,
      "learning_rate": 9.842723492616913e-06,
      "loss": 0.4785,
      "num_input_tokens_seen": 18386488,
      "step": 28110
    },
    {
      "epoch": 14.735324947589099,
      "grad_norm": 0.27349400520324707,
      "learning_rate": 9.833632045251207e-06,
      "loss": 0.3196,
      "num_input_tokens_seen": 18389272,
      "step": 28115
    },
    {
      "epoch": 14.737945492662474,
      "grad_norm": 0.6116654276847839,
      "learning_rate": 9.824543770516656e-06,
      "loss": 0.3823,
      "num_input_tokens_seen": 18393240,
      "step": 28120
    },
    {
      "epoch": 14.74056603773585,
      "grad_norm": 0.5204229354858398,
      "learning_rate": 9.815458670314412e-06,
      "loss": 0.4364,
      "num_input_tokens_seen": 18396504,
      "step": 28125
    },
    {
      "epoch": 14.743186582809225,
      "grad_norm": 0.9417074918746948,
      "learning_rate": 9.80637674654498e-06,
      "loss": 0.3803,
      "num_input_tokens_seen": 18399096,
      "step": 28130
    },
    {
      "epoch": 14.7458071278826,
      "grad_norm": 0.570692777633667,
      "learning_rate": 9.797298001108193e-06,
      "loss": 0.4327,
      "num_input_tokens_seen": 18402936,
      "step": 28135
    },
    {
      "epoch": 14.748427672955975,
      "grad_norm": 0.693198025226593,
      "learning_rate": 9.788222435903232e-06,
      "loss": 0.437,
      "num_input_tokens_seen": 18406008,
      "step": 28140
    },
    {
      "epoch": 14.75104821802935,
      "grad_norm": 0.2993186116218567,
      "learning_rate": 9.779150052828596e-06,
      "loss": 0.4952,
      "num_input_tokens_seen": 18409752,
      "step": 28145
    },
    {
      "epoch": 14.753668763102725,
      "grad_norm": 0.6799495816230774,
      "learning_rate": 9.770080853782108e-06,
      "loss": 0.3743,
      "num_input_tokens_seen": 18412408,
      "step": 28150
    },
    {
      "epoch": 14.7562893081761,
      "grad_norm": 0.7034666538238525,
      "learning_rate": 9.76101484066095e-06,
      "loss": 0.5993,
      "num_input_tokens_seen": 18415672,
      "step": 28155
    },
    {
      "epoch": 14.758909853249476,
      "grad_norm": 0.42790400981903076,
      "learning_rate": 9.751952015361637e-06,
      "loss": 0.5606,
      "num_input_tokens_seen": 18418904,
      "step": 28160
    },
    {
      "epoch": 14.76153039832285,
      "grad_norm": 0.3553830683231354,
      "learning_rate": 9.742892379780003e-06,
      "loss": 0.4136,
      "num_input_tokens_seen": 18421816,
      "step": 28165
    },
    {
      "epoch": 14.764150943396226,
      "grad_norm": 0.6073037385940552,
      "learning_rate": 9.733835935811237e-06,
      "loss": 0.6253,
      "num_input_tokens_seen": 18424696,
      "step": 28170
    },
    {
      "epoch": 14.766771488469601,
      "grad_norm": 0.71804279088974,
      "learning_rate": 9.724782685349809e-06,
      "loss": 0.4263,
      "num_input_tokens_seen": 18427864,
      "step": 28175
    },
    {
      "epoch": 14.769392033542976,
      "grad_norm": 0.6358601450920105,
      "learning_rate": 9.715732630289576e-06,
      "loss": 0.602,
      "num_input_tokens_seen": 18430872,
      "step": 28180
    },
    {
      "epoch": 14.772012578616351,
      "grad_norm": 0.6548779606819153,
      "learning_rate": 9.706685772523701e-06,
      "loss": 0.4075,
      "num_input_tokens_seen": 18433400,
      "step": 28185
    },
    {
      "epoch": 14.774633123689728,
      "grad_norm": 0.41700828075408936,
      "learning_rate": 9.697642113944692e-06,
      "loss": 0.5351,
      "num_input_tokens_seen": 18436536,
      "step": 28190
    },
    {
      "epoch": 14.777253668763104,
      "grad_norm": 0.32233986258506775,
      "learning_rate": 9.688601656444381e-06,
      "loss": 0.3511,
      "num_input_tokens_seen": 18439896,
      "step": 28195
    },
    {
      "epoch": 14.779874213836479,
      "grad_norm": 0.3323477506637573,
      "learning_rate": 9.679564401913921e-06,
      "loss": 0.4078,
      "num_input_tokens_seen": 18445112,
      "step": 28200
    },
    {
      "epoch": 14.782494758909854,
      "grad_norm": 0.69271320104599,
      "learning_rate": 9.670530352243798e-06,
      "loss": 0.3529,
      "num_input_tokens_seen": 18448056,
      "step": 28205
    },
    {
      "epoch": 14.785115303983229,
      "grad_norm": 1.0538105964660645,
      "learning_rate": 9.661499509323835e-06,
      "loss": 0.4042,
      "num_input_tokens_seen": 18451320,
      "step": 28210
    },
    {
      "epoch": 14.787735849056604,
      "grad_norm": 0.27892592549324036,
      "learning_rate": 9.652471875043182e-06,
      "loss": 0.3358,
      "num_input_tokens_seen": 18454168,
      "step": 28215
    },
    {
      "epoch": 14.79035639412998,
      "grad_norm": 0.47376540303230286,
      "learning_rate": 9.64344745129033e-06,
      "loss": 0.4764,
      "num_input_tokens_seen": 18457144,
      "step": 28220
    },
    {
      "epoch": 14.792976939203355,
      "grad_norm": 0.32627546787261963,
      "learning_rate": 9.634426239953073e-06,
      "loss": 0.431,
      "num_input_tokens_seen": 18460088,
      "step": 28225
    },
    {
      "epoch": 14.79559748427673,
      "grad_norm": 0.8147494196891785,
      "learning_rate": 9.62540824291854e-06,
      "loss": 0.5122,
      "num_input_tokens_seen": 18463800,
      "step": 28230
    },
    {
      "epoch": 14.798218029350105,
      "grad_norm": 0.30577725172042847,
      "learning_rate": 9.616393462073197e-06,
      "loss": 0.5275,
      "num_input_tokens_seen": 18466744,
      "step": 28235
    },
    {
      "epoch": 14.80083857442348,
      "grad_norm": 0.8290197849273682,
      "learning_rate": 9.607381899302839e-06,
      "loss": 0.5905,
      "num_input_tokens_seen": 18469848,
      "step": 28240
    },
    {
      "epoch": 14.803459119496855,
      "grad_norm": 0.42599329352378845,
      "learning_rate": 9.59837355649258e-06,
      "loss": 0.5326,
      "num_input_tokens_seen": 18473496,
      "step": 28245
    },
    {
      "epoch": 14.80607966457023,
      "grad_norm": 0.6564911007881165,
      "learning_rate": 9.589368435526868e-06,
      "loss": 0.4484,
      "num_input_tokens_seen": 18477048,
      "step": 28250
    },
    {
      "epoch": 14.808700209643606,
      "grad_norm": 0.5980159044265747,
      "learning_rate": 9.580366538289468e-06,
      "loss": 0.4916,
      "num_input_tokens_seen": 18479640,
      "step": 28255
    },
    {
      "epoch": 14.81132075471698,
      "grad_norm": 0.40995410084724426,
      "learning_rate": 9.57136786666346e-06,
      "loss": 0.4745,
      "num_input_tokens_seen": 18482808,
      "step": 28260
    },
    {
      "epoch": 14.813941299790356,
      "grad_norm": 0.6268957257270813,
      "learning_rate": 9.562372422531276e-06,
      "loss": 0.5212,
      "num_input_tokens_seen": 18485496,
      "step": 28265
    },
    {
      "epoch": 14.816561844863731,
      "grad_norm": 0.3121080994606018,
      "learning_rate": 9.553380207774662e-06,
      "loss": 0.4007,
      "num_input_tokens_seen": 18488568,
      "step": 28270
    },
    {
      "epoch": 14.819182389937106,
      "grad_norm": 0.4533627927303314,
      "learning_rate": 9.54439122427469e-06,
      "loss": 0.4404,
      "num_input_tokens_seen": 18491256,
      "step": 28275
    },
    {
      "epoch": 14.821802935010481,
      "grad_norm": 0.6610106825828552,
      "learning_rate": 9.535405473911748e-06,
      "loss": 0.3977,
      "num_input_tokens_seen": 18494424,
      "step": 28280
    },
    {
      "epoch": 14.824423480083858,
      "grad_norm": 0.35323482751846313,
      "learning_rate": 9.526422958565546e-06,
      "loss": 0.4548,
      "num_input_tokens_seen": 18497496,
      "step": 28285
    },
    {
      "epoch": 14.827044025157234,
      "grad_norm": 0.4937571883201599,
      "learning_rate": 9.517443680115124e-06,
      "loss": 0.4143,
      "num_input_tokens_seen": 18501304,
      "step": 28290
    },
    {
      "epoch": 14.829664570230609,
      "grad_norm": 0.7490524053573608,
      "learning_rate": 9.50846764043885e-06,
      "loss": 0.4422,
      "num_input_tokens_seen": 18505048,
      "step": 28295
    },
    {
      "epoch": 14.832285115303984,
      "grad_norm": 0.5236074924468994,
      "learning_rate": 9.499494841414422e-06,
      "loss": 0.4792,
      "num_input_tokens_seen": 18508216,
      "step": 28300
    },
    {
      "epoch": 14.834905660377359,
      "grad_norm": 0.32095372676849365,
      "learning_rate": 9.490525284918822e-06,
      "loss": 0.4391,
      "num_input_tokens_seen": 18511192,
      "step": 28305
    },
    {
      "epoch": 14.837526205450734,
      "grad_norm": 0.7939698696136475,
      "learning_rate": 9.4815589728284e-06,
      "loss": 0.4119,
      "num_input_tokens_seen": 18514296,
      "step": 28310
    },
    {
      "epoch": 14.84014675052411,
      "grad_norm": 0.34144723415374756,
      "learning_rate": 9.472595907018788e-06,
      "loss": 0.5956,
      "num_input_tokens_seen": 18517560,
      "step": 28315
    },
    {
      "epoch": 14.842767295597485,
      "grad_norm": 0.8341229557991028,
      "learning_rate": 9.463636089364968e-06,
      "loss": 0.4855,
      "num_input_tokens_seen": 18520888,
      "step": 28320
    },
    {
      "epoch": 14.84538784067086,
      "grad_norm": 0.44531622529029846,
      "learning_rate": 9.454679521741239e-06,
      "loss": 0.4651,
      "num_input_tokens_seen": 18525016,
      "step": 28325
    },
    {
      "epoch": 14.848008385744235,
      "grad_norm": 0.3940308392047882,
      "learning_rate": 9.445726206021194e-06,
      "loss": 0.514,
      "num_input_tokens_seen": 18529112,
      "step": 28330
    },
    {
      "epoch": 14.85062893081761,
      "grad_norm": 0.27582040429115295,
      "learning_rate": 9.436776144077772e-06,
      "loss": 0.4506,
      "num_input_tokens_seen": 18532344,
      "step": 28335
    },
    {
      "epoch": 14.853249475890985,
      "grad_norm": 0.4611615836620331,
      "learning_rate": 9.427829337783237e-06,
      "loss": 0.4284,
      "num_input_tokens_seen": 18535608,
      "step": 28340
    },
    {
      "epoch": 14.85587002096436,
      "grad_norm": 0.6009258031845093,
      "learning_rate": 9.418885789009138e-06,
      "loss": 0.4177,
      "num_input_tokens_seen": 18538200,
      "step": 28345
    },
    {
      "epoch": 14.858490566037736,
      "grad_norm": 0.5449191927909851,
      "learning_rate": 9.409945499626374e-06,
      "loss": 0.3281,
      "num_input_tokens_seen": 18540600,
      "step": 28350
    },
    {
      "epoch": 14.86111111111111,
      "grad_norm": 0.5371561646461487,
      "learning_rate": 9.40100847150516e-06,
      "loss": 0.4503,
      "num_input_tokens_seen": 18543192,
      "step": 28355
    },
    {
      "epoch": 14.863731656184486,
      "grad_norm": 0.7728439569473267,
      "learning_rate": 9.392074706515003e-06,
      "loss": 0.4581,
      "num_input_tokens_seen": 18546488,
      "step": 28360
    },
    {
      "epoch": 14.866352201257861,
      "grad_norm": 0.34494203329086304,
      "learning_rate": 9.383144206524753e-06,
      "loss": 0.485,
      "num_input_tokens_seen": 18549752,
      "step": 28365
    },
    {
      "epoch": 14.868972746331236,
      "grad_norm": 0.6612276434898376,
      "learning_rate": 9.374216973402575e-06,
      "loss": 0.4272,
      "num_input_tokens_seen": 18552696,
      "step": 28370
    },
    {
      "epoch": 14.871593291404611,
      "grad_norm": 1.0472062826156616,
      "learning_rate": 9.365293009015926e-06,
      "loss": 0.5157,
      "num_input_tokens_seen": 18555416,
      "step": 28375
    },
    {
      "epoch": 14.874213836477988,
      "grad_norm": 0.597852349281311,
      "learning_rate": 9.356372315231623e-06,
      "loss": 0.6467,
      "num_input_tokens_seen": 18558712,
      "step": 28380
    },
    {
      "epoch": 14.876834381551364,
      "grad_norm": 0.4765341877937317,
      "learning_rate": 9.347454893915749e-06,
      "loss": 0.4161,
      "num_input_tokens_seen": 18562072,
      "step": 28385
    },
    {
      "epoch": 14.879454926624739,
      "grad_norm": 0.6596824526786804,
      "learning_rate": 9.338540746933741e-06,
      "loss": 0.5236,
      "num_input_tokens_seen": 18565208,
      "step": 28390
    },
    {
      "epoch": 14.882075471698114,
      "grad_norm": 0.47083780169487,
      "learning_rate": 9.329629876150339e-06,
      "loss": 0.4753,
      "num_input_tokens_seen": 18568568,
      "step": 28395
    },
    {
      "epoch": 14.884696016771489,
      "grad_norm": 0.5973259806632996,
      "learning_rate": 9.320722283429581e-06,
      "loss": 0.3951,
      "num_input_tokens_seen": 18571384,
      "step": 28400
    },
    {
      "epoch": 14.887316561844864,
      "grad_norm": 0.5007336735725403,
      "learning_rate": 9.311817970634854e-06,
      "loss": 0.3092,
      "num_input_tokens_seen": 18574680,
      "step": 28405
    },
    {
      "epoch": 14.88993710691824,
      "grad_norm": 0.42553046345710754,
      "learning_rate": 9.302916939628814e-06,
      "loss": 0.3739,
      "num_input_tokens_seen": 18577464,
      "step": 28410
    },
    {
      "epoch": 14.892557651991615,
      "grad_norm": 0.29728350043296814,
      "learning_rate": 9.29401919227347e-06,
      "loss": 0.3675,
      "num_input_tokens_seen": 18580024,
      "step": 28415
    },
    {
      "epoch": 14.89517819706499,
      "grad_norm": 0.9792747497558594,
      "learning_rate": 9.28512473043013e-06,
      "loss": 0.4185,
      "num_input_tokens_seen": 18582840,
      "step": 28420
    },
    {
      "epoch": 14.897798742138365,
      "grad_norm": 0.5356178879737854,
      "learning_rate": 9.276233555959419e-06,
      "loss": 0.6327,
      "num_input_tokens_seen": 18585880,
      "step": 28425
    },
    {
      "epoch": 14.90041928721174,
      "grad_norm": 0.43451008200645447,
      "learning_rate": 9.267345670721254e-06,
      "loss": 0.4501,
      "num_input_tokens_seen": 18588920,
      "step": 28430
    },
    {
      "epoch": 14.903039832285115,
      "grad_norm": 0.6458855867385864,
      "learning_rate": 9.258461076574895e-06,
      "loss": 0.3896,
      "num_input_tokens_seen": 18591736,
      "step": 28435
    },
    {
      "epoch": 14.90566037735849,
      "grad_norm": 1.087475061416626,
      "learning_rate": 9.24957977537888e-06,
      "loss": 0.5305,
      "num_input_tokens_seen": 18594392,
      "step": 28440
    },
    {
      "epoch": 14.908280922431866,
      "grad_norm": 0.2827858328819275,
      "learning_rate": 9.240701768991086e-06,
      "loss": 0.4905,
      "num_input_tokens_seen": 18598424,
      "step": 28445
    },
    {
      "epoch": 14.91090146750524,
      "grad_norm": 0.5645639300346375,
      "learning_rate": 9.231827059268692e-06,
      "loss": 0.3997,
      "num_input_tokens_seen": 18601976,
      "step": 28450
    },
    {
      "epoch": 14.913522012578616,
      "grad_norm": 1.247879981994629,
      "learning_rate": 9.222955648068193e-06,
      "loss": 0.6043,
      "num_input_tokens_seen": 18604568,
      "step": 28455
    },
    {
      "epoch": 14.916142557651991,
      "grad_norm": 0.48632436990737915,
      "learning_rate": 9.214087537245378e-06,
      "loss": 0.3541,
      "num_input_tokens_seen": 18607544,
      "step": 28460
    },
    {
      "epoch": 14.918763102725366,
      "grad_norm": 0.4976900517940521,
      "learning_rate": 9.205222728655344e-06,
      "loss": 0.3925,
      "num_input_tokens_seen": 18611352,
      "step": 28465
    },
    {
      "epoch": 14.921383647798741,
      "grad_norm": 0.18200145661830902,
      "learning_rate": 9.196361224152522e-06,
      "loss": 0.332,
      "num_input_tokens_seen": 18616920,
      "step": 28470
    },
    {
      "epoch": 14.924004192872118,
      "grad_norm": 0.8254332542419434,
      "learning_rate": 9.187503025590632e-06,
      "loss": 0.6722,
      "num_input_tokens_seen": 18619544,
      "step": 28475
    },
    {
      "epoch": 14.926624737945493,
      "grad_norm": 0.3203074634075165,
      "learning_rate": 9.17864813482272e-06,
      "loss": 0.3605,
      "num_input_tokens_seen": 18622072,
      "step": 28480
    },
    {
      "epoch": 14.929245283018869,
      "grad_norm": 0.4043824374675751,
      "learning_rate": 9.169796553701122e-06,
      "loss": 0.5962,
      "num_input_tokens_seen": 18624952,
      "step": 28485
    },
    {
      "epoch": 14.931865828092244,
      "grad_norm": 0.5405214428901672,
      "learning_rate": 9.160948284077473e-06,
      "loss": 0.635,
      "num_input_tokens_seen": 18627672,
      "step": 28490
    },
    {
      "epoch": 14.934486373165619,
      "grad_norm": 0.4031587243080139,
      "learning_rate": 9.152103327802738e-06,
      "loss": 0.4031,
      "num_input_tokens_seen": 18631160,
      "step": 28495
    },
    {
      "epoch": 14.937106918238994,
      "grad_norm": 0.657494843006134,
      "learning_rate": 9.143261686727189e-06,
      "loss": 0.4179,
      "num_input_tokens_seen": 18634168,
      "step": 28500
    },
    {
      "epoch": 14.93972746331237,
      "grad_norm": 0.5524824857711792,
      "learning_rate": 9.134423362700392e-06,
      "loss": 0.3711,
      "num_input_tokens_seen": 18636984,
      "step": 28505
    },
    {
      "epoch": 14.942348008385745,
      "grad_norm": 0.4239942133426666,
      "learning_rate": 9.12558835757123e-06,
      "loss": 0.483,
      "num_input_tokens_seen": 18639864,
      "step": 28510
    },
    {
      "epoch": 14.94496855345912,
      "grad_norm": 0.5305491089820862,
      "learning_rate": 9.116756673187878e-06,
      "loss": 0.4138,
      "num_input_tokens_seen": 18642264,
      "step": 28515
    },
    {
      "epoch": 14.947589098532495,
      "grad_norm": 0.30046167969703674,
      "learning_rate": 9.107928311397817e-06,
      "loss": 0.5034,
      "num_input_tokens_seen": 18646072,
      "step": 28520
    },
    {
      "epoch": 14.95020964360587,
      "grad_norm": 0.6420190930366516,
      "learning_rate": 9.09910327404785e-06,
      "loss": 0.3852,
      "num_input_tokens_seen": 18648696,
      "step": 28525
    },
    {
      "epoch": 14.952830188679245,
      "grad_norm": 0.33773621916770935,
      "learning_rate": 9.090281562984068e-06,
      "loss": 0.4135,
      "num_input_tokens_seen": 18651608,
      "step": 28530
    },
    {
      "epoch": 14.95545073375262,
      "grad_norm": 0.7263003587722778,
      "learning_rate": 9.081463180051883e-06,
      "loss": 0.4479,
      "num_input_tokens_seen": 18655064,
      "step": 28535
    },
    {
      "epoch": 14.958071278825996,
      "grad_norm": 0.6402848958969116,
      "learning_rate": 9.072648127095996e-06,
      "loss": 0.3691,
      "num_input_tokens_seen": 18658040,
      "step": 28540
    },
    {
      "epoch": 14.96069182389937,
      "grad_norm": 0.9051545858383179,
      "learning_rate": 9.0638364059604e-06,
      "loss": 0.4104,
      "num_input_tokens_seen": 18660472,
      "step": 28545
    },
    {
      "epoch": 14.963312368972746,
      "grad_norm": 0.3851456344127655,
      "learning_rate": 9.055028018488417e-06,
      "loss": 0.461,
      "num_input_tokens_seen": 18664216,
      "step": 28550
    },
    {
      "epoch": 14.965932914046121,
      "grad_norm": 0.30329227447509766,
      "learning_rate": 9.046222966522665e-06,
      "loss": 0.3324,
      "num_input_tokens_seen": 18667224,
      "step": 28555
    },
    {
      "epoch": 14.968553459119496,
      "grad_norm": 0.5122453570365906,
      "learning_rate": 9.037421251905059e-06,
      "loss": 0.4297,
      "num_input_tokens_seen": 18670008,
      "step": 28560
    },
    {
      "epoch": 14.971174004192871,
      "grad_norm": 0.5427181720733643,
      "learning_rate": 9.028622876476833e-06,
      "loss": 0.39,
      "num_input_tokens_seen": 18672888,
      "step": 28565
    },
    {
      "epoch": 14.973794549266248,
      "grad_norm": 0.4662100076675415,
      "learning_rate": 9.019827842078474e-06,
      "loss": 0.3794,
      "num_input_tokens_seen": 18676440,
      "step": 28570
    },
    {
      "epoch": 14.976415094339622,
      "grad_norm": 0.5643084645271301,
      "learning_rate": 9.011036150549817e-06,
      "loss": 0.4762,
      "num_input_tokens_seen": 18679480,
      "step": 28575
    },
    {
      "epoch": 14.979035639412999,
      "grad_norm": 0.8370704054832458,
      "learning_rate": 9.002247803729987e-06,
      "loss": 0.4417,
      "num_input_tokens_seen": 18683096,
      "step": 28580
    },
    {
      "epoch": 14.981656184486374,
      "grad_norm": 1.9784376621246338,
      "learning_rate": 8.993462803457404e-06,
      "loss": 0.5779,
      "num_input_tokens_seen": 18685304,
      "step": 28585
    },
    {
      "epoch": 14.984276729559749,
      "grad_norm": 0.569431722164154,
      "learning_rate": 8.984681151569799e-06,
      "loss": 0.5098,
      "num_input_tokens_seen": 18687640,
      "step": 28590
    },
    {
      "epoch": 14.986897274633124,
      "grad_norm": 0.5529800653457642,
      "learning_rate": 8.975902849904183e-06,
      "loss": 0.3562,
      "num_input_tokens_seen": 18691832,
      "step": 28595
    },
    {
      "epoch": 14.9895178197065,
      "grad_norm": 0.5079436302185059,
      "learning_rate": 8.967127900296874e-06,
      "loss": 0.6292,
      "num_input_tokens_seen": 18694744,
      "step": 28600
    },
    {
      "epoch": 14.992138364779874,
      "grad_norm": 0.3672895133495331,
      "learning_rate": 8.95835630458349e-06,
      "loss": 0.5401,
      "num_input_tokens_seen": 18699096,
      "step": 28605
    },
    {
      "epoch": 14.99475890985325,
      "grad_norm": 0.6337963938713074,
      "learning_rate": 8.949588064598958e-06,
      "loss": 0.552,
      "num_input_tokens_seen": 18702168,
      "step": 28610
    },
    {
      "epoch": 14.997379454926625,
      "grad_norm": 0.7380990386009216,
      "learning_rate": 8.940823182177499e-06,
      "loss": 0.4562,
      "num_input_tokens_seen": 18704920,
      "step": 28615
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.48919615149497986,
      "learning_rate": 8.932061659152607e-06,
      "loss": 0.3788,
      "num_input_tokens_seen": 18706976,
      "step": 28620
    },
    {
      "epoch": 15.0,
      "eval_loss": 0.4751826226711273,
      "eval_runtime": 9.8998,
      "eval_samples_per_second": 85.659,
      "eval_steps_per_second": 21.415,
      "num_input_tokens_seen": 18706976,
      "step": 28620
    },
    {
      "epoch": 15.002620545073375,
      "grad_norm": 0.5129375457763672,
      "learning_rate": 8.92330349735711e-06,
      "loss": 0.4015,
      "num_input_tokens_seen": 18710720,
      "step": 28625
    },
    {
      "epoch": 15.00524109014675,
      "grad_norm": 0.43807703256607056,
      "learning_rate": 8.914548698623104e-06,
      "loss": 0.4717,
      "num_input_tokens_seen": 18714528,
      "step": 28630
    },
    {
      "epoch": 15.007861635220126,
      "grad_norm": 0.8133870959281921,
      "learning_rate": 8.905797264781995e-06,
      "loss": 0.4048,
      "num_input_tokens_seen": 18717120,
      "step": 28635
    },
    {
      "epoch": 15.0104821802935,
      "grad_norm": 0.404352605342865,
      "learning_rate": 8.89704919766449e-06,
      "loss": 0.458,
      "num_input_tokens_seen": 18719776,
      "step": 28640
    },
    {
      "epoch": 15.013102725366876,
      "grad_norm": 0.3623303174972534,
      "learning_rate": 8.888304499100586e-06,
      "loss": 0.3974,
      "num_input_tokens_seen": 18722976,
      "step": 28645
    },
    {
      "epoch": 15.015723270440251,
      "grad_norm": 0.5630426406860352,
      "learning_rate": 8.879563170919566e-06,
      "loss": 0.5201,
      "num_input_tokens_seen": 18726432,
      "step": 28650
    },
    {
      "epoch": 15.018343815513626,
      "grad_norm": 0.5094785094261169,
      "learning_rate": 8.870825214950016e-06,
      "loss": 0.417,
      "num_input_tokens_seen": 18729440,
      "step": 28655
    },
    {
      "epoch": 15.020964360587001,
      "grad_norm": 0.5004783868789673,
      "learning_rate": 8.862090633019818e-06,
      "loss": 0.3808,
      "num_input_tokens_seen": 18732256,
      "step": 28660
    },
    {
      "epoch": 15.023584905660377,
      "grad_norm": 0.5173263549804688,
      "learning_rate": 8.853359426956146e-06,
      "loss": 0.4663,
      "num_input_tokens_seen": 18736480,
      "step": 28665
    },
    {
      "epoch": 15.026205450733753,
      "grad_norm": 1.232925534248352,
      "learning_rate": 8.844631598585482e-06,
      "loss": 0.4022,
      "num_input_tokens_seen": 18739200,
      "step": 28670
    },
    {
      "epoch": 15.028825995807129,
      "grad_norm": 0.7608414888381958,
      "learning_rate": 8.835907149733569e-06,
      "loss": 0.3232,
      "num_input_tokens_seen": 18741536,
      "step": 28675
    },
    {
      "epoch": 15.031446540880504,
      "grad_norm": 0.4494103789329529,
      "learning_rate": 8.827186082225477e-06,
      "loss": 0.5798,
      "num_input_tokens_seen": 18744672,
      "step": 28680
    },
    {
      "epoch": 15.034067085953879,
      "grad_norm": 1.126928687095642,
      "learning_rate": 8.81846839788554e-06,
      "loss": 0.5764,
      "num_input_tokens_seen": 18747904,
      "step": 28685
    },
    {
      "epoch": 15.036687631027254,
      "grad_norm": 0.3803693652153015,
      "learning_rate": 8.809754098537407e-06,
      "loss": 0.4401,
      "num_input_tokens_seen": 18752000,
      "step": 28690
    },
    {
      "epoch": 15.03930817610063,
      "grad_norm": 0.42881935834884644,
      "learning_rate": 8.801043186004016e-06,
      "loss": 0.4359,
      "num_input_tokens_seen": 18755616,
      "step": 28695
    },
    {
      "epoch": 15.041928721174004,
      "grad_norm": 0.875939667224884,
      "learning_rate": 8.792335662107575e-06,
      "loss": 0.5202,
      "num_input_tokens_seen": 18758240,
      "step": 28700
    },
    {
      "epoch": 15.04454926624738,
      "grad_norm": 0.4046550691127777,
      "learning_rate": 8.783631528669605e-06,
      "loss": 0.2928,
      "num_input_tokens_seen": 18762368,
      "step": 28705
    },
    {
      "epoch": 15.047169811320755,
      "grad_norm": 0.5587854981422424,
      "learning_rate": 8.774930787510924e-06,
      "loss": 0.4343,
      "num_input_tokens_seen": 18765152,
      "step": 28710
    },
    {
      "epoch": 15.04979035639413,
      "grad_norm": 0.4502461552619934,
      "learning_rate": 8.766233440451608e-06,
      "loss": 0.6357,
      "num_input_tokens_seen": 18768928,
      "step": 28715
    },
    {
      "epoch": 15.052410901467505,
      "grad_norm": 0.5875999927520752,
      "learning_rate": 8.75753948931105e-06,
      "loss": 0.3406,
      "num_input_tokens_seen": 18771488,
      "step": 28720
    },
    {
      "epoch": 15.05503144654088,
      "grad_norm": 0.5603253841400146,
      "learning_rate": 8.748848935907936e-06,
      "loss": 0.5125,
      "num_input_tokens_seen": 18774272,
      "step": 28725
    },
    {
      "epoch": 15.057651991614255,
      "grad_norm": 0.5604993104934692,
      "learning_rate": 8.740161782060214e-06,
      "loss": 0.5357,
      "num_input_tokens_seen": 18776704,
      "step": 28730
    },
    {
      "epoch": 15.06027253668763,
      "grad_norm": 0.7307776808738708,
      "learning_rate": 8.731478029585144e-06,
      "loss": 0.4984,
      "num_input_tokens_seen": 18779520,
      "step": 28735
    },
    {
      "epoch": 15.062893081761006,
      "grad_norm": 0.4837489426136017,
      "learning_rate": 8.722797680299278e-06,
      "loss": 0.4244,
      "num_input_tokens_seen": 18782336,
      "step": 28740
    },
    {
      "epoch": 15.065513626834381,
      "grad_norm": 0.4504036009311676,
      "learning_rate": 8.71412073601843e-06,
      "loss": 0.4608,
      "num_input_tokens_seen": 18785504,
      "step": 28745
    },
    {
      "epoch": 15.068134171907756,
      "grad_norm": 0.35241878032684326,
      "learning_rate": 8.705447198557731e-06,
      "loss": 0.3199,
      "num_input_tokens_seen": 18788480,
      "step": 28750
    },
    {
      "epoch": 15.070754716981131,
      "grad_norm": 0.4264705777168274,
      "learning_rate": 8.696777069731574e-06,
      "loss": 0.4978,
      "num_input_tokens_seen": 18794144,
      "step": 28755
    },
    {
      "epoch": 15.073375262054507,
      "grad_norm": 0.3494444489479065,
      "learning_rate": 8.688110351353654e-06,
      "loss": 0.4372,
      "num_input_tokens_seen": 18797376,
      "step": 28760
    },
    {
      "epoch": 15.075995807127883,
      "grad_norm": 0.4092191159725189,
      "learning_rate": 8.679447045236962e-06,
      "loss": 0.5038,
      "num_input_tokens_seen": 18800672,
      "step": 28765
    },
    {
      "epoch": 15.078616352201259,
      "grad_norm": 0.40287891030311584,
      "learning_rate": 8.670787153193746e-06,
      "loss": 0.4096,
      "num_input_tokens_seen": 18803648,
      "step": 28770
    },
    {
      "epoch": 15.081236897274634,
      "grad_norm": 0.7845495939254761,
      "learning_rate": 8.662130677035574e-06,
      "loss": 0.4311,
      "num_input_tokens_seen": 18806144,
      "step": 28775
    },
    {
      "epoch": 15.083857442348009,
      "grad_norm": 0.4697816073894501,
      "learning_rate": 8.65347761857326e-06,
      "loss": 0.4338,
      "num_input_tokens_seen": 18809152,
      "step": 28780
    },
    {
      "epoch": 15.086477987421384,
      "grad_norm": 0.818145215511322,
      "learning_rate": 8.64482797961694e-06,
      "loss": 0.4497,
      "num_input_tokens_seen": 18812192,
      "step": 28785
    },
    {
      "epoch": 15.08909853249476,
      "grad_norm": 0.673061728477478,
      "learning_rate": 8.636181761976016e-06,
      "loss": 0.4825,
      "num_input_tokens_seen": 18815168,
      "step": 28790
    },
    {
      "epoch": 15.091719077568134,
      "grad_norm": 0.618874728679657,
      "learning_rate": 8.62753896745919e-06,
      "loss": 0.3889,
      "num_input_tokens_seen": 18819072,
      "step": 28795
    },
    {
      "epoch": 15.09433962264151,
      "grad_norm": 0.6313915252685547,
      "learning_rate": 8.618899597874413e-06,
      "loss": 0.5882,
      "num_input_tokens_seen": 18821728,
      "step": 28800
    },
    {
      "epoch": 15.096960167714885,
      "grad_norm": 0.3910234570503235,
      "learning_rate": 8.610263655028964e-06,
      "loss": 0.4968,
      "num_input_tokens_seen": 18824544,
      "step": 28805
    },
    {
      "epoch": 15.09958071278826,
      "grad_norm": 0.4889160096645355,
      "learning_rate": 8.601631140729366e-06,
      "loss": 0.3236,
      "num_input_tokens_seen": 18828032,
      "step": 28810
    },
    {
      "epoch": 15.102201257861635,
      "grad_norm": 0.6480652093887329,
      "learning_rate": 8.593002056781451e-06,
      "loss": 0.4449,
      "num_input_tokens_seen": 18830688,
      "step": 28815
    },
    {
      "epoch": 15.10482180293501,
      "grad_norm": 0.42761340737342834,
      "learning_rate": 8.584376404990326e-06,
      "loss": 0.309,
      "num_input_tokens_seen": 18833664,
      "step": 28820
    },
    {
      "epoch": 15.107442348008385,
      "grad_norm": 0.29579800367355347,
      "learning_rate": 8.575754187160384e-06,
      "loss": 0.3533,
      "num_input_tokens_seen": 18837536,
      "step": 28825
    },
    {
      "epoch": 15.11006289308176,
      "grad_norm": 0.5838510394096375,
      "learning_rate": 8.567135405095291e-06,
      "loss": 0.4095,
      "num_input_tokens_seen": 18840000,
      "step": 28830
    },
    {
      "epoch": 15.112683438155136,
      "grad_norm": 0.2891126275062561,
      "learning_rate": 8.558520060597985e-06,
      "loss": 0.4743,
      "num_input_tokens_seen": 18844224,
      "step": 28835
    },
    {
      "epoch": 15.115303983228511,
      "grad_norm": 0.3610062301158905,
      "learning_rate": 8.54990815547071e-06,
      "loss": 0.3536,
      "num_input_tokens_seen": 18847104,
      "step": 28840
    },
    {
      "epoch": 15.117924528301886,
      "grad_norm": 0.7624406814575195,
      "learning_rate": 8.541299691514974e-06,
      "loss": 0.5728,
      "num_input_tokens_seen": 18849536,
      "step": 28845
    },
    {
      "epoch": 15.120545073375261,
      "grad_norm": 0.5131867527961731,
      "learning_rate": 8.53269467053158e-06,
      "loss": 0.406,
      "num_input_tokens_seen": 18854528,
      "step": 28850
    },
    {
      "epoch": 15.123165618448636,
      "grad_norm": 0.7380588054656982,
      "learning_rate": 8.524093094320593e-06,
      "loss": 0.4385,
      "num_input_tokens_seen": 18857344,
      "step": 28855
    },
    {
      "epoch": 15.125786163522013,
      "grad_norm": 0.8491824269294739,
      "learning_rate": 8.515494964681357e-06,
      "loss": 0.4411,
      "num_input_tokens_seen": 18859872,
      "step": 28860
    },
    {
      "epoch": 15.128406708595389,
      "grad_norm": 0.6414725184440613,
      "learning_rate": 8.506900283412506e-06,
      "loss": 0.2999,
      "num_input_tokens_seen": 18862240,
      "step": 28865
    },
    {
      "epoch": 15.131027253668764,
      "grad_norm": 0.3933183252811432,
      "learning_rate": 8.498309052311953e-06,
      "loss": 0.3524,
      "num_input_tokens_seen": 18865472,
      "step": 28870
    },
    {
      "epoch": 15.133647798742139,
      "grad_norm": 0.3480696380138397,
      "learning_rate": 8.489721273176887e-06,
      "loss": 0.4956,
      "num_input_tokens_seen": 18868960,
      "step": 28875
    },
    {
      "epoch": 15.136268343815514,
      "grad_norm": 0.49043533205986023,
      "learning_rate": 8.481136947803777e-06,
      "loss": 0.4473,
      "num_input_tokens_seen": 18871616,
      "step": 28880
    },
    {
      "epoch": 15.13888888888889,
      "grad_norm": 0.34059280157089233,
      "learning_rate": 8.472556077988362e-06,
      "loss": 0.5342,
      "num_input_tokens_seen": 18875040,
      "step": 28885
    },
    {
      "epoch": 15.141509433962264,
      "grad_norm": 0.6228618025779724,
      "learning_rate": 8.46397866552565e-06,
      "loss": 0.4779,
      "num_input_tokens_seen": 18877632,
      "step": 28890
    },
    {
      "epoch": 15.14412997903564,
      "grad_norm": 0.4513218104839325,
      "learning_rate": 8.455404712209952e-06,
      "loss": 0.4777,
      "num_input_tokens_seen": 18881056,
      "step": 28895
    },
    {
      "epoch": 15.146750524109015,
      "grad_norm": 1.3060383796691895,
      "learning_rate": 8.446834219834836e-06,
      "loss": 0.5014,
      "num_input_tokens_seen": 18883616,
      "step": 28900
    },
    {
      "epoch": 15.14937106918239,
      "grad_norm": 0.46982645988464355,
      "learning_rate": 8.43826719019316e-06,
      "loss": 0.4055,
      "num_input_tokens_seen": 18886592,
      "step": 28905
    },
    {
      "epoch": 15.151991614255765,
      "grad_norm": 0.3781747817993164,
      "learning_rate": 8.429703625077043e-06,
      "loss": 0.4171,
      "num_input_tokens_seen": 18891104,
      "step": 28910
    },
    {
      "epoch": 15.15461215932914,
      "grad_norm": 0.3764469623565674,
      "learning_rate": 8.421143526277875e-06,
      "loss": 0.3409,
      "num_input_tokens_seen": 18894528,
      "step": 28915
    },
    {
      "epoch": 15.157232704402515,
      "grad_norm": 0.46568533778190613,
      "learning_rate": 8.41258689558634e-06,
      "loss": 0.4922,
      "num_input_tokens_seen": 18897760,
      "step": 28920
    },
    {
      "epoch": 15.15985324947589,
      "grad_norm": 0.28819388151168823,
      "learning_rate": 8.404033734792386e-06,
      "loss": 0.4267,
      "num_input_tokens_seen": 18900800,
      "step": 28925
    },
    {
      "epoch": 15.162473794549266,
      "grad_norm": 0.508269190788269,
      "learning_rate": 8.395484045685248e-06,
      "loss": 0.4323,
      "num_input_tokens_seen": 18903680,
      "step": 28930
    },
    {
      "epoch": 15.165094339622641,
      "grad_norm": 0.5607662796974182,
      "learning_rate": 8.386937830053412e-06,
      "loss": 0.4396,
      "num_input_tokens_seen": 18906336,
      "step": 28935
    },
    {
      "epoch": 15.167714884696016,
      "grad_norm": 0.5193312764167786,
      "learning_rate": 8.37839508968464e-06,
      "loss": 0.46,
      "num_input_tokens_seen": 18910112,
      "step": 28940
    },
    {
      "epoch": 15.170335429769391,
      "grad_norm": 0.7765271067619324,
      "learning_rate": 8.369855826365988e-06,
      "loss": 0.4636,
      "num_input_tokens_seen": 18912928,
      "step": 28945
    },
    {
      "epoch": 15.172955974842766,
      "grad_norm": 0.37182432413101196,
      "learning_rate": 8.361320041883772e-06,
      "loss": 0.4756,
      "num_input_tokens_seen": 18916416,
      "step": 28950
    },
    {
      "epoch": 15.175576519916143,
      "grad_norm": 0.5161542296409607,
      "learning_rate": 8.352787738023576e-06,
      "loss": 0.5726,
      "num_input_tokens_seen": 18919968,
      "step": 28955
    },
    {
      "epoch": 15.178197064989519,
      "grad_norm": 0.39816907048225403,
      "learning_rate": 8.344258916570274e-06,
      "loss": 0.356,
      "num_input_tokens_seen": 18923392,
      "step": 28960
    },
    {
      "epoch": 15.180817610062894,
      "grad_norm": 0.3638218641281128,
      "learning_rate": 8.335733579307988e-06,
      "loss": 0.4274,
      "num_input_tokens_seen": 18926528,
      "step": 28965
    },
    {
      "epoch": 15.183438155136269,
      "grad_norm": 0.3936798870563507,
      "learning_rate": 8.327211728020113e-06,
      "loss": 0.3175,
      "num_input_tokens_seen": 18929696,
      "step": 28970
    },
    {
      "epoch": 15.186058700209644,
      "grad_norm": 0.4986741542816162,
      "learning_rate": 8.31869336448933e-06,
      "loss": 0.4242,
      "num_input_tokens_seen": 18932384,
      "step": 28975
    },
    {
      "epoch": 15.18867924528302,
      "grad_norm": 0.396084189414978,
      "learning_rate": 8.310178490497586e-06,
      "loss": 0.4446,
      "num_input_tokens_seen": 18934880,
      "step": 28980
    },
    {
      "epoch": 15.191299790356394,
      "grad_norm": 0.6556990146636963,
      "learning_rate": 8.301667107826103e-06,
      "loss": 0.3554,
      "num_input_tokens_seen": 18938816,
      "step": 28985
    },
    {
      "epoch": 15.19392033542977,
      "grad_norm": 0.6845309734344482,
      "learning_rate": 8.293159218255345e-06,
      "loss": 0.4152,
      "num_input_tokens_seen": 18941824,
      "step": 28990
    },
    {
      "epoch": 15.196540880503145,
      "grad_norm": 0.8847886919975281,
      "learning_rate": 8.284654823565088e-06,
      "loss": 0.433,
      "num_input_tokens_seen": 18945056,
      "step": 28995
    },
    {
      "epoch": 15.19916142557652,
      "grad_norm": 0.5071887969970703,
      "learning_rate": 8.276153925534333e-06,
      "loss": 0.5007,
      "num_input_tokens_seen": 18948256,
      "step": 29000
    },
    {
      "epoch": 15.201781970649895,
      "grad_norm": 0.749905526638031,
      "learning_rate": 8.267656525941383e-06,
      "loss": 0.5016,
      "num_input_tokens_seen": 18951008,
      "step": 29005
    },
    {
      "epoch": 15.20440251572327,
      "grad_norm": 0.459812730550766,
      "learning_rate": 8.259162626563801e-06,
      "loss": 0.4097,
      "num_input_tokens_seen": 18954464,
      "step": 29010
    },
    {
      "epoch": 15.207023060796645,
      "grad_norm": 0.26780587434768677,
      "learning_rate": 8.250672229178402e-06,
      "loss": 0.4156,
      "num_input_tokens_seen": 18958464,
      "step": 29015
    },
    {
      "epoch": 15.20964360587002,
      "grad_norm": 0.4630773663520813,
      "learning_rate": 8.242185335561295e-06,
      "loss": 0.4327,
      "num_input_tokens_seen": 18961152,
      "step": 29020
    },
    {
      "epoch": 15.212264150943396,
      "grad_norm": 0.4432715177536011,
      "learning_rate": 8.23370194748782e-06,
      "loss": 0.3142,
      "num_input_tokens_seen": 18964640,
      "step": 29025
    },
    {
      "epoch": 15.214884696016771,
      "grad_norm": 0.4390435218811035,
      "learning_rate": 8.22522206673262e-06,
      "loss": 0.4282,
      "num_input_tokens_seen": 18967616,
      "step": 29030
    },
    {
      "epoch": 15.217505241090146,
      "grad_norm": 0.4076436161994934,
      "learning_rate": 8.216745695069589e-06,
      "loss": 0.4808,
      "num_input_tokens_seen": 18971136,
      "step": 29035
    },
    {
      "epoch": 15.220125786163521,
      "grad_norm": 0.414289265871048,
      "learning_rate": 8.208272834271894e-06,
      "loss": 0.5747,
      "num_input_tokens_seen": 18974208,
      "step": 29040
    },
    {
      "epoch": 15.222746331236896,
      "grad_norm": 0.46489012241363525,
      "learning_rate": 8.19980348611194e-06,
      "loss": 0.4449,
      "num_input_tokens_seen": 18977408,
      "step": 29045
    },
    {
      "epoch": 15.225366876310272,
      "grad_norm": 0.7775939702987671,
      "learning_rate": 8.191337652361439e-06,
      "loss": 0.4901,
      "num_input_tokens_seen": 18980288,
      "step": 29050
    },
    {
      "epoch": 15.227987421383649,
      "grad_norm": 0.5572308897972107,
      "learning_rate": 8.182875334791332e-06,
      "loss": 0.3412,
      "num_input_tokens_seen": 18984096,
      "step": 29055
    },
    {
      "epoch": 15.230607966457024,
      "grad_norm": 0.7978036999702454,
      "learning_rate": 8.174416535171841e-06,
      "loss": 0.4606,
      "num_input_tokens_seen": 18987648,
      "step": 29060
    },
    {
      "epoch": 15.233228511530399,
      "grad_norm": 0.3043300211429596,
      "learning_rate": 8.165961255272467e-06,
      "loss": 0.3937,
      "num_input_tokens_seen": 18990208,
      "step": 29065
    },
    {
      "epoch": 15.235849056603774,
      "grad_norm": 0.5588286519050598,
      "learning_rate": 8.157509496861935e-06,
      "loss": 0.39,
      "num_input_tokens_seen": 18993440,
      "step": 29070
    },
    {
      "epoch": 15.23846960167715,
      "grad_norm": 0.43279147148132324,
      "learning_rate": 8.149061261708266e-06,
      "loss": 0.4193,
      "num_input_tokens_seen": 18997024,
      "step": 29075
    },
    {
      "epoch": 15.241090146750524,
      "grad_norm": 1.0962885618209839,
      "learning_rate": 8.140616551578745e-06,
      "loss": 0.4518,
      "num_input_tokens_seen": 19000864,
      "step": 29080
    },
    {
      "epoch": 15.2437106918239,
      "grad_norm": 0.36062759160995483,
      "learning_rate": 8.13217536823989e-06,
      "loss": 0.4675,
      "num_input_tokens_seen": 19004576,
      "step": 29085
    },
    {
      "epoch": 15.246331236897275,
      "grad_norm": 0.7384195327758789,
      "learning_rate": 8.12373771345752e-06,
      "loss": 0.4039,
      "num_input_tokens_seen": 19008128,
      "step": 29090
    },
    {
      "epoch": 15.24895178197065,
      "grad_norm": 0.5126553177833557,
      "learning_rate": 8.115303588996676e-06,
      "loss": 0.4926,
      "num_input_tokens_seen": 19013952,
      "step": 29095
    },
    {
      "epoch": 15.251572327044025,
      "grad_norm": 0.4853196442127228,
      "learning_rate": 8.10687299662169e-06,
      "loss": 0.5064,
      "num_input_tokens_seen": 19017216,
      "step": 29100
    },
    {
      "epoch": 15.2541928721174,
      "grad_norm": 0.36759239435195923,
      "learning_rate": 8.098445938096147e-06,
      "loss": 0.3689,
      "num_input_tokens_seen": 19020096,
      "step": 29105
    },
    {
      "epoch": 15.256813417190775,
      "grad_norm": 0.5403875112533569,
      "learning_rate": 8.090022415182898e-06,
      "loss": 0.4117,
      "num_input_tokens_seen": 19023456,
      "step": 29110
    },
    {
      "epoch": 15.25943396226415,
      "grad_norm": 0.610575795173645,
      "learning_rate": 8.081602429644033e-06,
      "loss": 0.4756,
      "num_input_tokens_seen": 19027424,
      "step": 29115
    },
    {
      "epoch": 15.262054507337526,
      "grad_norm": 0.7132812142372131,
      "learning_rate": 8.073185983240932e-06,
      "loss": 0.5175,
      "num_input_tokens_seen": 19030976,
      "step": 29120
    },
    {
      "epoch": 15.264675052410901,
      "grad_norm": 0.5393431782722473,
      "learning_rate": 8.064773077734206e-06,
      "loss": 0.4141,
      "num_input_tokens_seen": 19034976,
      "step": 29125
    },
    {
      "epoch": 15.267295597484276,
      "grad_norm": 0.3534494638442993,
      "learning_rate": 8.05636371488374e-06,
      "loss": 0.4353,
      "num_input_tokens_seen": 19038752,
      "step": 29130
    },
    {
      "epoch": 15.269916142557651,
      "grad_norm": 0.4879676401615143,
      "learning_rate": 8.047957896448696e-06,
      "loss": 0.4615,
      "num_input_tokens_seen": 19041536,
      "step": 29135
    },
    {
      "epoch": 15.272536687631026,
      "grad_norm": 0.548579216003418,
      "learning_rate": 8.039555624187451e-06,
      "loss": 0.4307,
      "num_input_tokens_seen": 19044448,
      "step": 29140
    },
    {
      "epoch": 15.275157232704402,
      "grad_norm": 0.6061315536499023,
      "learning_rate": 8.031156899857681e-06,
      "loss": 0.5141,
      "num_input_tokens_seen": 19047904,
      "step": 29145
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 0.2619558870792389,
      "learning_rate": 8.022761725216288e-06,
      "loss": 0.5064,
      "num_input_tokens_seen": 19052864,
      "step": 29150
    },
    {
      "epoch": 15.280398322851154,
      "grad_norm": 0.3811267018318176,
      "learning_rate": 8.014370102019456e-06,
      "loss": 0.5865,
      "num_input_tokens_seen": 19055456,
      "step": 29155
    },
    {
      "epoch": 15.283018867924529,
      "grad_norm": 0.576664388179779,
      "learning_rate": 8.005982032022616e-06,
      "loss": 0.3813,
      "num_input_tokens_seen": 19058240,
      "step": 29160
    },
    {
      "epoch": 15.285639412997904,
      "grad_norm": 0.36462321877479553,
      "learning_rate": 7.997597516980467e-06,
      "loss": 0.4543,
      "num_input_tokens_seen": 19062336,
      "step": 29165
    },
    {
      "epoch": 15.28825995807128,
      "grad_norm": 0.41808652877807617,
      "learning_rate": 7.989216558646942e-06,
      "loss": 0.2911,
      "num_input_tokens_seen": 19066592,
      "step": 29170
    },
    {
      "epoch": 15.290880503144654,
      "grad_norm": 0.33758988976478577,
      "learning_rate": 7.980839158775235e-06,
      "loss": 0.3978,
      "num_input_tokens_seen": 19069888,
      "step": 29175
    },
    {
      "epoch": 15.29350104821803,
      "grad_norm": 0.415577232837677,
      "learning_rate": 7.972465319117814e-06,
      "loss": 0.3519,
      "num_input_tokens_seen": 19073152,
      "step": 29180
    },
    {
      "epoch": 15.296121593291405,
      "grad_norm": 0.4176323115825653,
      "learning_rate": 7.964095041426386e-06,
      "loss": 0.3753,
      "num_input_tokens_seen": 19075904,
      "step": 29185
    },
    {
      "epoch": 15.29874213836478,
      "grad_norm": 0.4003797471523285,
      "learning_rate": 7.95572832745192e-06,
      "loss": 0.6033,
      "num_input_tokens_seen": 19079584,
      "step": 29190
    },
    {
      "epoch": 15.301362683438155,
      "grad_norm": 0.37236011028289795,
      "learning_rate": 7.947365178944643e-06,
      "loss": 0.3651,
      "num_input_tokens_seen": 19082656,
      "step": 29195
    },
    {
      "epoch": 15.30398322851153,
      "grad_norm": 0.29038676619529724,
      "learning_rate": 7.939005597654025e-06,
      "loss": 0.4714,
      "num_input_tokens_seen": 19086208,
      "step": 29200
    },
    {
      "epoch": 15.306603773584905,
      "grad_norm": 0.42842942476272583,
      "learning_rate": 7.930649585328787e-06,
      "loss": 0.4014,
      "num_input_tokens_seen": 19089184,
      "step": 29205
    },
    {
      "epoch": 15.30922431865828,
      "grad_norm": 0.4657934308052063,
      "learning_rate": 7.922297143716918e-06,
      "loss": 0.4253,
      "num_input_tokens_seen": 19093664,
      "step": 29210
    },
    {
      "epoch": 15.311844863731656,
      "grad_norm": 0.27428218722343445,
      "learning_rate": 7.913948274565652e-06,
      "loss": 0.3857,
      "num_input_tokens_seen": 19098048,
      "step": 29215
    },
    {
      "epoch": 15.314465408805031,
      "grad_norm": 0.4299681484699249,
      "learning_rate": 7.905602979621491e-06,
      "loss": 0.4316,
      "num_input_tokens_seen": 19101760,
      "step": 29220
    },
    {
      "epoch": 15.317085953878406,
      "grad_norm": 0.6426909565925598,
      "learning_rate": 7.897261260630159e-06,
      "loss": 0.3807,
      "num_input_tokens_seen": 19106912,
      "step": 29225
    },
    {
      "epoch": 15.319706498951781,
      "grad_norm": 0.31673434376716614,
      "learning_rate": 7.888923119336647e-06,
      "loss": 0.495,
      "num_input_tokens_seen": 19110272,
      "step": 29230
    },
    {
      "epoch": 15.322327044025156,
      "grad_norm": 0.5348682999610901,
      "learning_rate": 7.880588557485203e-06,
      "loss": 0.4524,
      "num_input_tokens_seen": 19113408,
      "step": 29235
    },
    {
      "epoch": 15.324947589098532,
      "grad_norm": 0.37410613894462585,
      "learning_rate": 7.872257576819325e-06,
      "loss": 0.5187,
      "num_input_tokens_seen": 19116864,
      "step": 29240
    },
    {
      "epoch": 15.327568134171909,
      "grad_norm": 0.2844918370246887,
      "learning_rate": 7.863930179081769e-06,
      "loss": 0.3785,
      "num_input_tokens_seen": 19120576,
      "step": 29245
    },
    {
      "epoch": 15.330188679245284,
      "grad_norm": 0.7903721928596497,
      "learning_rate": 7.855606366014517e-06,
      "loss": 0.4182,
      "num_input_tokens_seen": 19123872,
      "step": 29250
    },
    {
      "epoch": 15.332809224318659,
      "grad_norm": 0.4560420513153076,
      "learning_rate": 7.847286139358814e-06,
      "loss": 0.395,
      "num_input_tokens_seen": 19126688,
      "step": 29255
    },
    {
      "epoch": 15.335429769392034,
      "grad_norm": 0.4669332206249237,
      "learning_rate": 7.838969500855162e-06,
      "loss": 0.3307,
      "num_input_tokens_seen": 19129504,
      "step": 29260
    },
    {
      "epoch": 15.33805031446541,
      "grad_norm": 0.7493042349815369,
      "learning_rate": 7.830656452243307e-06,
      "loss": 0.5981,
      "num_input_tokens_seen": 19132064,
      "step": 29265
    },
    {
      "epoch": 15.340670859538784,
      "grad_norm": 0.4201829731464386,
      "learning_rate": 7.822346995262241e-06,
      "loss": 0.2934,
      "num_input_tokens_seen": 19134848,
      "step": 29270
    },
    {
      "epoch": 15.34329140461216,
      "grad_norm": 0.42553162574768066,
      "learning_rate": 7.814041131650223e-06,
      "loss": 0.6358,
      "num_input_tokens_seen": 19137632,
      "step": 29275
    },
    {
      "epoch": 15.345911949685535,
      "grad_norm": 0.44089746475219727,
      "learning_rate": 7.805738863144731e-06,
      "loss": 0.4212,
      "num_input_tokens_seen": 19140672,
      "step": 29280
    },
    {
      "epoch": 15.34853249475891,
      "grad_norm": 0.2780754864215851,
      "learning_rate": 7.797440191482496e-06,
      "loss": 0.4686,
      "num_input_tokens_seen": 19144192,
      "step": 29285
    },
    {
      "epoch": 15.351153039832285,
      "grad_norm": 0.7114400267601013,
      "learning_rate": 7.789145118399518e-06,
      "loss": 0.4586,
      "num_input_tokens_seen": 19146880,
      "step": 29290
    },
    {
      "epoch": 15.35377358490566,
      "grad_norm": 0.4014926850795746,
      "learning_rate": 7.780853645631032e-06,
      "loss": 0.4117,
      "num_input_tokens_seen": 19150400,
      "step": 29295
    },
    {
      "epoch": 15.356394129979035,
      "grad_norm": 0.6618772149085999,
      "learning_rate": 7.772565774911522e-06,
      "loss": 0.3822,
      "num_input_tokens_seen": 19153504,
      "step": 29300
    },
    {
      "epoch": 15.35901467505241,
      "grad_norm": 0.3113057017326355,
      "learning_rate": 7.76428150797471e-06,
      "loss": 0.3835,
      "num_input_tokens_seen": 19156960,
      "step": 29305
    },
    {
      "epoch": 15.361635220125786,
      "grad_norm": 0.7469022274017334,
      "learning_rate": 7.756000846553562e-06,
      "loss": 0.6785,
      "num_input_tokens_seen": 19159744,
      "step": 29310
    },
    {
      "epoch": 15.364255765199161,
      "grad_norm": 0.44393569231033325,
      "learning_rate": 7.74772379238031e-06,
      "loss": 0.3872,
      "num_input_tokens_seen": 19163808,
      "step": 29315
    },
    {
      "epoch": 15.366876310272536,
      "grad_norm": 0.7724894881248474,
      "learning_rate": 7.739450347186417e-06,
      "loss": 0.4197,
      "num_input_tokens_seen": 19166304,
      "step": 29320
    },
    {
      "epoch": 15.369496855345911,
      "grad_norm": 0.37221840023994446,
      "learning_rate": 7.731180512702599e-06,
      "loss": 0.409,
      "num_input_tokens_seen": 19169792,
      "step": 29325
    },
    {
      "epoch": 15.372117400419286,
      "grad_norm": 0.3300062417984009,
      "learning_rate": 7.722914290658795e-06,
      "loss": 0.4643,
      "num_input_tokens_seen": 19172448,
      "step": 29330
    },
    {
      "epoch": 15.374737945492662,
      "grad_norm": 0.43126678466796875,
      "learning_rate": 7.714651682784224e-06,
      "loss": 0.3929,
      "num_input_tokens_seen": 19176352,
      "step": 29335
    },
    {
      "epoch": 15.377358490566039,
      "grad_norm": 1.1582129001617432,
      "learning_rate": 7.706392690807313e-06,
      "loss": 0.5752,
      "num_input_tokens_seen": 19179808,
      "step": 29340
    },
    {
      "epoch": 15.379979035639414,
      "grad_norm": 0.5824013948440552,
      "learning_rate": 7.69813731645575e-06,
      "loss": 0.4309,
      "num_input_tokens_seen": 19183872,
      "step": 29345
    },
    {
      "epoch": 15.382599580712789,
      "grad_norm": 0.37602415680885315,
      "learning_rate": 7.689885561456475e-06,
      "loss": 0.3719,
      "num_input_tokens_seen": 19187584,
      "step": 29350
    },
    {
      "epoch": 15.385220125786164,
      "grad_norm": 0.36085206270217896,
      "learning_rate": 7.681637427535663e-06,
      "loss": 0.4743,
      "num_input_tokens_seen": 19192320,
      "step": 29355
    },
    {
      "epoch": 15.38784067085954,
      "grad_norm": 0.4443141520023346,
      "learning_rate": 7.673392916418715e-06,
      "loss": 0.4173,
      "num_input_tokens_seen": 19195712,
      "step": 29360
    },
    {
      "epoch": 15.390461215932914,
      "grad_norm": 0.42652377486228943,
      "learning_rate": 7.665152029830303e-06,
      "loss": 0.5466,
      "num_input_tokens_seen": 19199328,
      "step": 29365
    },
    {
      "epoch": 15.39308176100629,
      "grad_norm": 0.6720343232154846,
      "learning_rate": 7.656914769494314e-06,
      "loss": 0.455,
      "num_input_tokens_seen": 19202368,
      "step": 29370
    },
    {
      "epoch": 15.395702306079665,
      "grad_norm": 0.3265088200569153,
      "learning_rate": 7.648681137133892e-06,
      "loss": 0.4987,
      "num_input_tokens_seen": 19205728,
      "step": 29375
    },
    {
      "epoch": 15.39832285115304,
      "grad_norm": 0.5727311968803406,
      "learning_rate": 7.640451134471432e-06,
      "loss": 0.3474,
      "num_input_tokens_seen": 19208768,
      "step": 29380
    },
    {
      "epoch": 15.400943396226415,
      "grad_norm": 0.49877896904945374,
      "learning_rate": 7.632224763228538e-06,
      "loss": 0.5159,
      "num_input_tokens_seen": 19213152,
      "step": 29385
    },
    {
      "epoch": 15.40356394129979,
      "grad_norm": 0.3721107542514801,
      "learning_rate": 7.6240020251260896e-06,
      "loss": 0.4985,
      "num_input_tokens_seen": 19215392,
      "step": 29390
    },
    {
      "epoch": 15.406184486373165,
      "grad_norm": 0.3335210382938385,
      "learning_rate": 7.615782921884174e-06,
      "loss": 0.444,
      "num_input_tokens_seen": 19218944,
      "step": 29395
    },
    {
      "epoch": 15.40880503144654,
      "grad_norm": 0.4198457896709442,
      "learning_rate": 7.607567455222142e-06,
      "loss": 0.5348,
      "num_input_tokens_seen": 19221984,
      "step": 29400
    },
    {
      "epoch": 15.411425576519916,
      "grad_norm": 0.37082144618034363,
      "learning_rate": 7.599355626858582e-06,
      "loss": 0.5822,
      "num_input_tokens_seen": 19225376,
      "step": 29405
    },
    {
      "epoch": 15.414046121593291,
      "grad_norm": 0.5163974761962891,
      "learning_rate": 7.591147438511298e-06,
      "loss": 0.4785,
      "num_input_tokens_seen": 19228448,
      "step": 29410
    },
    {
      "epoch": 15.416666666666666,
      "grad_norm": 0.40933582186698914,
      "learning_rate": 7.5829428918973625e-06,
      "loss": 0.761,
      "num_input_tokens_seen": 19231712,
      "step": 29415
    },
    {
      "epoch": 15.419287211740041,
      "grad_norm": 0.33761826157569885,
      "learning_rate": 7.574741988733075e-06,
      "loss": 0.516,
      "num_input_tokens_seen": 19234144,
      "step": 29420
    },
    {
      "epoch": 15.421907756813416,
      "grad_norm": 0.41854995489120483,
      "learning_rate": 7.56654473073396e-06,
      "loss": 0.4924,
      "num_input_tokens_seen": 19237056,
      "step": 29425
    },
    {
      "epoch": 15.424528301886792,
      "grad_norm": 0.36226341128349304,
      "learning_rate": 7.558351119614796e-06,
      "loss": 0.4089,
      "num_input_tokens_seen": 19239904,
      "step": 29430
    },
    {
      "epoch": 15.427148846960169,
      "grad_norm": 0.414002388715744,
      "learning_rate": 7.5501611570896e-06,
      "loss": 0.4475,
      "num_input_tokens_seen": 19243552,
      "step": 29435
    },
    {
      "epoch": 15.429769392033544,
      "grad_norm": 0.4801689386367798,
      "learning_rate": 7.541974844871602e-06,
      "loss": 0.4598,
      "num_input_tokens_seen": 19246560,
      "step": 29440
    },
    {
      "epoch": 15.432389937106919,
      "grad_norm": 0.24215756356716156,
      "learning_rate": 7.5337921846732965e-06,
      "loss": 0.4374,
      "num_input_tokens_seen": 19249632,
      "step": 29445
    },
    {
      "epoch": 15.435010482180294,
      "grad_norm": 0.6564838290214539,
      "learning_rate": 7.525613178206409e-06,
      "loss": 0.4456,
      "num_input_tokens_seen": 19252736,
      "step": 29450
    },
    {
      "epoch": 15.43763102725367,
      "grad_norm": 0.3776251971721649,
      "learning_rate": 7.517437827181878e-06,
      "loss": 0.4168,
      "num_input_tokens_seen": 19255968,
      "step": 29455
    },
    {
      "epoch": 15.440251572327044,
      "grad_norm": 0.5057605504989624,
      "learning_rate": 7.509266133309908e-06,
      "loss": 0.4703,
      "num_input_tokens_seen": 19259680,
      "step": 29460
    },
    {
      "epoch": 15.44287211740042,
      "grad_norm": 0.5269046425819397,
      "learning_rate": 7.501098098299914e-06,
      "loss": 0.4069,
      "num_input_tokens_seen": 19262848,
      "step": 29465
    },
    {
      "epoch": 15.445492662473795,
      "grad_norm": 0.5279726386070251,
      "learning_rate": 7.492933723860557e-06,
      "loss": 0.4074,
      "num_input_tokens_seen": 19265536,
      "step": 29470
    },
    {
      "epoch": 15.44811320754717,
      "grad_norm": 0.6616098284721375,
      "learning_rate": 7.4847730116997334e-06,
      "loss": 0.3416,
      "num_input_tokens_seen": 19268384,
      "step": 29475
    },
    {
      "epoch": 15.450733752620545,
      "grad_norm": 0.21621422469615936,
      "learning_rate": 7.476615963524583e-06,
      "loss": 0.439,
      "num_input_tokens_seen": 19272000,
      "step": 29480
    },
    {
      "epoch": 15.45335429769392,
      "grad_norm": 0.421470582485199,
      "learning_rate": 7.468462581041452e-06,
      "loss": 0.5206,
      "num_input_tokens_seen": 19275360,
      "step": 29485
    },
    {
      "epoch": 15.455974842767295,
      "grad_norm": 0.586388111114502,
      "learning_rate": 7.460312865955932e-06,
      "loss": 0.4416,
      "num_input_tokens_seen": 19279488,
      "step": 29490
    },
    {
      "epoch": 15.45859538784067,
      "grad_norm": 0.9730382561683655,
      "learning_rate": 7.4521668199728584e-06,
      "loss": 0.4496,
      "num_input_tokens_seen": 19282176,
      "step": 29495
    },
    {
      "epoch": 15.461215932914046,
      "grad_norm": 0.6478029489517212,
      "learning_rate": 7.4440244447962884e-06,
      "loss": 0.5257,
      "num_input_tokens_seen": 19285056,
      "step": 29500
    },
    {
      "epoch": 15.463836477987421,
      "grad_norm": 0.41865333914756775,
      "learning_rate": 7.435885742129523e-06,
      "loss": 0.4062,
      "num_input_tokens_seen": 19288704,
      "step": 29505
    },
    {
      "epoch": 15.466457023060796,
      "grad_norm": 0.5178580284118652,
      "learning_rate": 7.427750713675071e-06,
      "loss": 0.6868,
      "num_input_tokens_seen": 19291584,
      "step": 29510
    },
    {
      "epoch": 15.469077568134171,
      "grad_norm": 0.7442618012428284,
      "learning_rate": 7.419619361134702e-06,
      "loss": 0.3784,
      "num_input_tokens_seen": 19295072,
      "step": 29515
    },
    {
      "epoch": 15.471698113207546,
      "grad_norm": 0.48504409193992615,
      "learning_rate": 7.411491686209387e-06,
      "loss": 0.4315,
      "num_input_tokens_seen": 19298144,
      "step": 29520
    },
    {
      "epoch": 15.474318658280922,
      "grad_norm": 0.5011841654777527,
      "learning_rate": 7.4033676905993525e-06,
      "loss": 0.6073,
      "num_input_tokens_seen": 19301472,
      "step": 29525
    },
    {
      "epoch": 15.476939203354299,
      "grad_norm": 0.6772369146347046,
      "learning_rate": 7.3952473760040444e-06,
      "loss": 0.4347,
      "num_input_tokens_seen": 19303936,
      "step": 29530
    },
    {
      "epoch": 15.479559748427674,
      "grad_norm": 0.30701449513435364,
      "learning_rate": 7.387130744122148e-06,
      "loss": 0.4694,
      "num_input_tokens_seen": 19308416,
      "step": 29535
    },
    {
      "epoch": 15.482180293501049,
      "grad_norm": 0.8789471983909607,
      "learning_rate": 7.379017796651558e-06,
      "loss": 0.4906,
      "num_input_tokens_seen": 19312992,
      "step": 29540
    },
    {
      "epoch": 15.484800838574424,
      "grad_norm": 0.994197428226471,
      "learning_rate": 7.3709085352894085e-06,
      "loss": 0.3263,
      "num_input_tokens_seen": 19317792,
      "step": 29545
    },
    {
      "epoch": 15.4874213836478,
      "grad_norm": 0.6327720880508423,
      "learning_rate": 7.362802961732071e-06,
      "loss": 0.3925,
      "num_input_tokens_seen": 19320928,
      "step": 29550
    },
    {
      "epoch": 15.490041928721174,
      "grad_norm": 0.752507209777832,
      "learning_rate": 7.354701077675136e-06,
      "loss": 0.4773,
      "num_input_tokens_seen": 19323584,
      "step": 29555
    },
    {
      "epoch": 15.49266247379455,
      "grad_norm": 0.39320600032806396,
      "learning_rate": 7.346602884813439e-06,
      "loss": 0.4285,
      "num_input_tokens_seen": 19326112,
      "step": 29560
    },
    {
      "epoch": 15.495283018867925,
      "grad_norm": 0.35707008838653564,
      "learning_rate": 7.338508384841014e-06,
      "loss": 0.4234,
      "num_input_tokens_seen": 19329824,
      "step": 29565
    },
    {
      "epoch": 15.4979035639413,
      "grad_norm": 0.4748323857784271,
      "learning_rate": 7.3304175794511345e-06,
      "loss": 0.4606,
      "num_input_tokens_seen": 19332544,
      "step": 29570
    },
    {
      "epoch": 15.500524109014675,
      "grad_norm": 0.6312762498855591,
      "learning_rate": 7.3223304703363135e-06,
      "loss": 0.3866,
      "num_input_tokens_seen": 19335744,
      "step": 29575
    },
    {
      "epoch": 15.50314465408805,
      "grad_norm": 0.5023013949394226,
      "learning_rate": 7.314247059188281e-06,
      "loss": 0.5991,
      "num_input_tokens_seen": 19339456,
      "step": 29580
    },
    {
      "epoch": 15.505765199161425,
      "grad_norm": 0.5950062870979309,
      "learning_rate": 7.306167347697992e-06,
      "loss": 0.5597,
      "num_input_tokens_seen": 19342752,
      "step": 29585
    },
    {
      "epoch": 15.5083857442348,
      "grad_norm": 0.3532489240169525,
      "learning_rate": 7.29809133755564e-06,
      "loss": 0.4372,
      "num_input_tokens_seen": 19345536,
      "step": 29590
    },
    {
      "epoch": 15.511006289308176,
      "grad_norm": 0.678107738494873,
      "learning_rate": 7.290019030450629e-06,
      "loss": 0.4162,
      "num_input_tokens_seen": 19348768,
      "step": 29595
    },
    {
      "epoch": 15.51362683438155,
      "grad_norm": 0.4889757037162781,
      "learning_rate": 7.2819504280715794e-06,
      "loss": 0.3866,
      "num_input_tokens_seen": 19352128,
      "step": 29600
    },
    {
      "epoch": 15.516247379454926,
      "grad_norm": 0.3991105556488037,
      "learning_rate": 7.273885532106364e-06,
      "loss": 0.5239,
      "num_input_tokens_seen": 19354784,
      "step": 29605
    },
    {
      "epoch": 15.518867924528301,
      "grad_norm": 0.4085574150085449,
      "learning_rate": 7.265824344242064e-06,
      "loss": 0.5245,
      "num_input_tokens_seen": 19358528,
      "step": 29610
    },
    {
      "epoch": 15.521488469601676,
      "grad_norm": 0.6930716633796692,
      "learning_rate": 7.257766866165e-06,
      "loss": 0.3819,
      "num_input_tokens_seen": 19361760,
      "step": 29615
    },
    {
      "epoch": 15.524109014675052,
      "grad_norm": 0.3678540885448456,
      "learning_rate": 7.24971309956069e-06,
      "loss": 0.4115,
      "num_input_tokens_seen": 19365504,
      "step": 29620
    },
    {
      "epoch": 15.526729559748428,
      "grad_norm": 1.531760334968567,
      "learning_rate": 7.241663046113887e-06,
      "loss": 0.2982,
      "num_input_tokens_seen": 19369344,
      "step": 29625
    },
    {
      "epoch": 15.529350104821804,
      "grad_norm": 0.7110329866409302,
      "learning_rate": 7.233616707508576e-06,
      "loss": 0.4306,
      "num_input_tokens_seen": 19372224,
      "step": 29630
    },
    {
      "epoch": 15.531970649895179,
      "grad_norm": 0.6943836212158203,
      "learning_rate": 7.225574085427961e-06,
      "loss": 0.4175,
      "num_input_tokens_seen": 19374976,
      "step": 29635
    },
    {
      "epoch": 15.534591194968554,
      "grad_norm": 0.5506744980812073,
      "learning_rate": 7.217535181554474e-06,
      "loss": 0.3496,
      "num_input_tokens_seen": 19377440,
      "step": 29640
    },
    {
      "epoch": 15.53721174004193,
      "grad_norm": 0.7680844068527222,
      "learning_rate": 7.209499997569747e-06,
      "loss": 0.4483,
      "num_input_tokens_seen": 19380032,
      "step": 29645
    },
    {
      "epoch": 15.539832285115304,
      "grad_norm": 0.370352566242218,
      "learning_rate": 7.201468535154663e-06,
      "loss": 0.3389,
      "num_input_tokens_seen": 19383488,
      "step": 29650
    },
    {
      "epoch": 15.54245283018868,
      "grad_norm": 0.6711134910583496,
      "learning_rate": 7.193440795989295e-06,
      "loss": 0.5055,
      "num_input_tokens_seen": 19386560,
      "step": 29655
    },
    {
      "epoch": 15.545073375262055,
      "grad_norm": 0.48472777009010315,
      "learning_rate": 7.185416781752965e-06,
      "loss": 0.4895,
      "num_input_tokens_seen": 19389216,
      "step": 29660
    },
    {
      "epoch": 15.54769392033543,
      "grad_norm": 0.5032469630241394,
      "learning_rate": 7.177396494124206e-06,
      "loss": 0.4972,
      "num_input_tokens_seen": 19393056,
      "step": 29665
    },
    {
      "epoch": 15.550314465408805,
      "grad_norm": 0.6923936009407043,
      "learning_rate": 7.169379934780779e-06,
      "loss": 0.5344,
      "num_input_tokens_seen": 19396672,
      "step": 29670
    },
    {
      "epoch": 15.55293501048218,
      "grad_norm": 0.7936863899230957,
      "learning_rate": 7.161367105399644e-06,
      "loss": 0.5938,
      "num_input_tokens_seen": 19399168,
      "step": 29675
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 0.16914081573486328,
      "learning_rate": 7.153358007656991e-06,
      "loss": 0.3744,
      "num_input_tokens_seen": 19405248,
      "step": 29680
    },
    {
      "epoch": 15.55817610062893,
      "grad_norm": 0.6425229907035828,
      "learning_rate": 7.145352643228237e-06,
      "loss": 0.4527,
      "num_input_tokens_seen": 19407712,
      "step": 29685
    },
    {
      "epoch": 15.560796645702306,
      "grad_norm": 0.4653152525424957,
      "learning_rate": 7.137351013788013e-06,
      "loss": 0.3644,
      "num_input_tokens_seen": 19410816,
      "step": 29690
    },
    {
      "epoch": 15.56341719077568,
      "grad_norm": 0.4184160530567169,
      "learning_rate": 7.129353121010177e-06,
      "loss": 0.4065,
      "num_input_tokens_seen": 19413888,
      "step": 29695
    },
    {
      "epoch": 15.566037735849056,
      "grad_norm": 0.5700523257255554,
      "learning_rate": 7.121358966567779e-06,
      "loss": 0.4777,
      "num_input_tokens_seen": 19417344,
      "step": 29700
    },
    {
      "epoch": 15.568658280922431,
      "grad_norm": 0.5481350421905518,
      "learning_rate": 7.113368552133126e-06,
      "loss": 0.3306,
      "num_input_tokens_seen": 19420512,
      "step": 29705
    },
    {
      "epoch": 15.571278825995806,
      "grad_norm": 0.34705737233161926,
      "learning_rate": 7.105381879377701e-06,
      "loss": 0.3387,
      "num_input_tokens_seen": 19423328,
      "step": 29710
    },
    {
      "epoch": 15.573899371069182,
      "grad_norm": 0.44846439361572266,
      "learning_rate": 7.097398949972234e-06,
      "loss": 0.4483,
      "num_input_tokens_seen": 19427136,
      "step": 29715
    },
    {
      "epoch": 15.576519916142558,
      "grad_norm": 0.406951367855072,
      "learning_rate": 7.089419765586672e-06,
      "loss": 0.441,
      "num_input_tokens_seen": 19430496,
      "step": 29720
    },
    {
      "epoch": 15.579140461215934,
      "grad_norm": 0.6909473538398743,
      "learning_rate": 7.081444327890152e-06,
      "loss": 0.4446,
      "num_input_tokens_seen": 19433792,
      "step": 29725
    },
    {
      "epoch": 15.581761006289309,
      "grad_norm": 0.5311359763145447,
      "learning_rate": 7.073472638551054e-06,
      "loss": 0.3586,
      "num_input_tokens_seen": 19436384,
      "step": 29730
    },
    {
      "epoch": 15.584381551362684,
      "grad_norm": 0.7239390015602112,
      "learning_rate": 7.065504699236969e-06,
      "loss": 0.4815,
      "num_input_tokens_seen": 19439200,
      "step": 29735
    },
    {
      "epoch": 15.58700209643606,
      "grad_norm": 0.567293643951416,
      "learning_rate": 7.057540511614688e-06,
      "loss": 0.399,
      "num_input_tokens_seen": 19442400,
      "step": 29740
    },
    {
      "epoch": 15.589622641509434,
      "grad_norm": 0.511098325252533,
      "learning_rate": 7.049580077350232e-06,
      "loss": 0.4594,
      "num_input_tokens_seen": 19446464,
      "step": 29745
    },
    {
      "epoch": 15.59224318658281,
      "grad_norm": 0.6976357698440552,
      "learning_rate": 7.041623398108843e-06,
      "loss": 0.3108,
      "num_input_tokens_seen": 19449600,
      "step": 29750
    },
    {
      "epoch": 15.594863731656185,
      "grad_norm": 0.38210636377334595,
      "learning_rate": 7.033670475554949e-06,
      "loss": 0.3563,
      "num_input_tokens_seen": 19452608,
      "step": 29755
    },
    {
      "epoch": 15.59748427672956,
      "grad_norm": 0.2566131055355072,
      "learning_rate": 7.025721311352224e-06,
      "loss": 0.3887,
      "num_input_tokens_seen": 19456832,
      "step": 29760
    },
    {
      "epoch": 15.600104821802935,
      "grad_norm": 0.3583810031414032,
      "learning_rate": 7.017775907163546e-06,
      "loss": 0.3874,
      "num_input_tokens_seen": 19459136,
      "step": 29765
    },
    {
      "epoch": 15.60272536687631,
      "grad_norm": 0.4514758884906769,
      "learning_rate": 7.009834264650989e-06,
      "loss": 0.3634,
      "num_input_tokens_seen": 19463360,
      "step": 29770
    },
    {
      "epoch": 15.605345911949685,
      "grad_norm": 0.7782182097434998,
      "learning_rate": 7.001896385475867e-06,
      "loss": 0.3523,
      "num_input_tokens_seen": 19466432,
      "step": 29775
    },
    {
      "epoch": 15.60796645702306,
      "grad_norm": 0.26794615387916565,
      "learning_rate": 6.993962271298682e-06,
      "loss": 0.3544,
      "num_input_tokens_seen": 19470592,
      "step": 29780
    },
    {
      "epoch": 15.610587002096436,
      "grad_norm": 0.3699607253074646,
      "learning_rate": 6.986031923779166e-06,
      "loss": 0.3432,
      "num_input_tokens_seen": 19474496,
      "step": 29785
    },
    {
      "epoch": 15.61320754716981,
      "grad_norm": 0.31842556595802307,
      "learning_rate": 6.978105344576264e-06,
      "loss": 0.3881,
      "num_input_tokens_seen": 19477792,
      "step": 29790
    },
    {
      "epoch": 15.615828092243186,
      "grad_norm": 0.5195119976997375,
      "learning_rate": 6.970182535348111e-06,
      "loss": 0.4182,
      "num_input_tokens_seen": 19480736,
      "step": 29795
    },
    {
      "epoch": 15.618448637316561,
      "grad_norm": 0.23988282680511475,
      "learning_rate": 6.962263497752086e-06,
      "loss": 0.4702,
      "num_input_tokens_seen": 19485312,
      "step": 29800
    },
    {
      "epoch": 15.621069182389936,
      "grad_norm": 0.3506494462490082,
      "learning_rate": 6.9543482334447436e-06,
      "loss": 0.3462,
      "num_input_tokens_seen": 19488544,
      "step": 29805
    },
    {
      "epoch": 15.623689727463312,
      "grad_norm": 0.49000000953674316,
      "learning_rate": 6.946436744081875e-06,
      "loss": 0.3476,
      "num_input_tokens_seen": 19491744,
      "step": 29810
    },
    {
      "epoch": 15.626310272536688,
      "grad_norm": 0.4597093164920807,
      "learning_rate": 6.938529031318472e-06,
      "loss": 0.4579,
      "num_input_tokens_seen": 19495136,
      "step": 29815
    },
    {
      "epoch": 15.628930817610064,
      "grad_norm": 0.6140472292900085,
      "learning_rate": 6.9306250968087485e-06,
      "loss": 0.4343,
      "num_input_tokens_seen": 19498272,
      "step": 29820
    },
    {
      "epoch": 15.631551362683439,
      "grad_norm": 0.6291892528533936,
      "learning_rate": 6.922724942206102e-06,
      "loss": 0.4064,
      "num_input_tokens_seen": 19501792,
      "step": 29825
    },
    {
      "epoch": 15.634171907756814,
      "grad_norm": 0.621122419834137,
      "learning_rate": 6.914828569163167e-06,
      "loss": 0.5785,
      "num_input_tokens_seen": 19504640,
      "step": 29830
    },
    {
      "epoch": 15.63679245283019,
      "grad_norm": 0.4558289051055908,
      "learning_rate": 6.906935979331763e-06,
      "loss": 0.3556,
      "num_input_tokens_seen": 19508032,
      "step": 29835
    },
    {
      "epoch": 15.639412997903564,
      "grad_norm": 0.32550498843193054,
      "learning_rate": 6.8990471743629356e-06,
      "loss": 0.34,
      "num_input_tokens_seen": 19511008,
      "step": 29840
    },
    {
      "epoch": 15.64203354297694,
      "grad_norm": 0.4272511899471283,
      "learning_rate": 6.89116215590693e-06,
      "loss": 0.5582,
      "num_input_tokens_seen": 19514240,
      "step": 29845
    },
    {
      "epoch": 15.644654088050315,
      "grad_norm": 0.7468852996826172,
      "learning_rate": 6.8832809256132146e-06,
      "loss": 0.3797,
      "num_input_tokens_seen": 19517056,
      "step": 29850
    },
    {
      "epoch": 15.64727463312369,
      "grad_norm": 1.2128528356552124,
      "learning_rate": 6.875403485130444e-06,
      "loss": 0.3961,
      "num_input_tokens_seen": 19520256,
      "step": 29855
    },
    {
      "epoch": 15.649895178197065,
      "grad_norm": 0.7059171199798584,
      "learning_rate": 6.867529836106479e-06,
      "loss": 0.4055,
      "num_input_tokens_seen": 19522976,
      "step": 29860
    },
    {
      "epoch": 15.65251572327044,
      "grad_norm": 0.6423560976982117,
      "learning_rate": 6.8596599801884045e-06,
      "loss": 0.4019,
      "num_input_tokens_seen": 19526784,
      "step": 29865
    },
    {
      "epoch": 15.655136268343815,
      "grad_norm": 0.5142738223075867,
      "learning_rate": 6.851793919022509e-06,
      "loss": 0.3627,
      "num_input_tokens_seen": 19529920,
      "step": 29870
    },
    {
      "epoch": 15.65775681341719,
      "grad_norm": 0.5994335412979126,
      "learning_rate": 6.843931654254285e-06,
      "loss": 0.4669,
      "num_input_tokens_seen": 19532608,
      "step": 29875
    },
    {
      "epoch": 15.660377358490566,
      "grad_norm": 0.561363160610199,
      "learning_rate": 6.836073187528425e-06,
      "loss": 0.3977,
      "num_input_tokens_seen": 19535680,
      "step": 29880
    },
    {
      "epoch": 15.66299790356394,
      "grad_norm": 0.7092675566673279,
      "learning_rate": 6.828218520488821e-06,
      "loss": 0.3851,
      "num_input_tokens_seen": 19538144,
      "step": 29885
    },
    {
      "epoch": 15.665618448637316,
      "grad_norm": 0.37072843313217163,
      "learning_rate": 6.820367654778589e-06,
      "loss": 0.2787,
      "num_input_tokens_seen": 19541344,
      "step": 29890
    },
    {
      "epoch": 15.668238993710691,
      "grad_norm": 0.4581809937953949,
      "learning_rate": 6.812520592040039e-06,
      "loss": 0.4741,
      "num_input_tokens_seen": 19544640,
      "step": 29895
    },
    {
      "epoch": 15.670859538784066,
      "grad_norm": 0.3504672944545746,
      "learning_rate": 6.804677333914689e-06,
      "loss": 0.3039,
      "num_input_tokens_seen": 19547872,
      "step": 29900
    },
    {
      "epoch": 15.673480083857442,
      "grad_norm": 0.4087054431438446,
      "learning_rate": 6.796837882043261e-06,
      "loss": 0.3712,
      "num_input_tokens_seen": 19550880,
      "step": 29905
    },
    {
      "epoch": 15.676100628930818,
      "grad_norm": 0.5704782009124756,
      "learning_rate": 6.7890022380656785e-06,
      "loss": 0.6196,
      "num_input_tokens_seen": 19554112,
      "step": 29910
    },
    {
      "epoch": 15.678721174004194,
      "grad_norm": 0.430287629365921,
      "learning_rate": 6.781170403621056e-06,
      "loss": 0.3876,
      "num_input_tokens_seen": 19557760,
      "step": 29915
    },
    {
      "epoch": 15.681341719077569,
      "grad_norm": 0.8179377913475037,
      "learning_rate": 6.773342380347736e-06,
      "loss": 0.4561,
      "num_input_tokens_seen": 19561312,
      "step": 29920
    },
    {
      "epoch": 15.683962264150944,
      "grad_norm": 0.5052322745323181,
      "learning_rate": 6.765518169883248e-06,
      "loss": 0.5907,
      "num_input_tokens_seen": 19565248,
      "step": 29925
    },
    {
      "epoch": 15.68658280922432,
      "grad_norm": 0.9065332412719727,
      "learning_rate": 6.757697773864338e-06,
      "loss": 0.4987,
      "num_input_tokens_seen": 19567808,
      "step": 29930
    },
    {
      "epoch": 15.689203354297694,
      "grad_norm": 0.4234066605567932,
      "learning_rate": 6.749881193926932e-06,
      "loss": 0.5495,
      "num_input_tokens_seen": 19570880,
      "step": 29935
    },
    {
      "epoch": 15.69182389937107,
      "grad_norm": 0.6235284209251404,
      "learning_rate": 6.742068431706167e-06,
      "loss": 0.4229,
      "num_input_tokens_seen": 19573536,
      "step": 29940
    },
    {
      "epoch": 15.694444444444445,
      "grad_norm": 0.6836614012718201,
      "learning_rate": 6.734259488836386e-06,
      "loss": 0.4481,
      "num_input_tokens_seen": 19577408,
      "step": 29945
    },
    {
      "epoch": 15.69706498951782,
      "grad_norm": 0.4451095461845398,
      "learning_rate": 6.726454366951135e-06,
      "loss": 0.4114,
      "num_input_tokens_seen": 19580960,
      "step": 29950
    },
    {
      "epoch": 15.699685534591195,
      "grad_norm": 0.6647276878356934,
      "learning_rate": 6.7186530676831616e-06,
      "loss": 0.6843,
      "num_input_tokens_seen": 19584448,
      "step": 29955
    },
    {
      "epoch": 15.70230607966457,
      "grad_norm": 0.34029167890548706,
      "learning_rate": 6.710855592664403e-06,
      "loss": 0.434,
      "num_input_tokens_seen": 19587392,
      "step": 29960
    },
    {
      "epoch": 15.704926624737945,
      "grad_norm": 1.1094982624053955,
      "learning_rate": 6.703061943525993e-06,
      "loss": 0.3721,
      "num_input_tokens_seen": 19590240,
      "step": 29965
    },
    {
      "epoch": 15.70754716981132,
      "grad_norm": 0.4753890931606293,
      "learning_rate": 6.69527212189828e-06,
      "loss": 0.406,
      "num_input_tokens_seen": 19593056,
      "step": 29970
    },
    {
      "epoch": 15.710167714884696,
      "grad_norm": 0.40664681792259216,
      "learning_rate": 6.687486129410811e-06,
      "loss": 0.3908,
      "num_input_tokens_seen": 19596160,
      "step": 29975
    },
    {
      "epoch": 15.71278825995807,
      "grad_norm": 0.6882717609405518,
      "learning_rate": 6.679703967692322e-06,
      "loss": 0.3148,
      "num_input_tokens_seen": 19601184,
      "step": 29980
    },
    {
      "epoch": 15.715408805031446,
      "grad_norm": 0.3282395899295807,
      "learning_rate": 6.67192563837076e-06,
      "loss": 0.3921,
      "num_input_tokens_seen": 19604160,
      "step": 29985
    },
    {
      "epoch": 15.718029350104821,
      "grad_norm": 0.3511640429496765,
      "learning_rate": 6.664151143073258e-06,
      "loss": 0.3656,
      "num_input_tokens_seen": 19607872,
      "step": 29990
    },
    {
      "epoch": 15.720649895178196,
      "grad_norm": 0.4594636559486389,
      "learning_rate": 6.656380483426141e-06,
      "loss": 0.3248,
      "num_input_tokens_seen": 19612096,
      "step": 29995
    },
    {
      "epoch": 15.723270440251572,
      "grad_norm": 0.5032331347465515,
      "learning_rate": 6.648613661054956e-06,
      "loss": 0.4528,
      "num_input_tokens_seen": 19614976,
      "step": 30000
    },
    {
      "epoch": 15.725890985324948,
      "grad_norm": 0.4605109989643097,
      "learning_rate": 6.6408506775844256e-06,
      "loss": 0.4799,
      "num_input_tokens_seen": 19618624,
      "step": 30005
    },
    {
      "epoch": 15.728511530398324,
      "grad_norm": 0.7876706719398499,
      "learning_rate": 6.633091534638492e-06,
      "loss": 0.5315,
      "num_input_tokens_seen": 19621920,
      "step": 30010
    },
    {
      "epoch": 15.731132075471699,
      "grad_norm": 0.36202481389045715,
      "learning_rate": 6.625336233840257e-06,
      "loss": 0.3754,
      "num_input_tokens_seen": 19624352,
      "step": 30015
    },
    {
      "epoch": 15.733752620545074,
      "grad_norm": 0.3523208498954773,
      "learning_rate": 6.617584776812064e-06,
      "loss": 0.3702,
      "num_input_tokens_seen": 19631744,
      "step": 30020
    },
    {
      "epoch": 15.73637316561845,
      "grad_norm": 0.6986294388771057,
      "learning_rate": 6.6098371651754085e-06,
      "loss": 0.3946,
      "num_input_tokens_seen": 19634592,
      "step": 30025
    },
    {
      "epoch": 15.738993710691824,
      "grad_norm": 0.4068385362625122,
      "learning_rate": 6.602093400551012e-06,
      "loss": 0.4509,
      "num_input_tokens_seen": 19637472,
      "step": 30030
    },
    {
      "epoch": 15.7416142557652,
      "grad_norm": 0.3721371591091156,
      "learning_rate": 6.59435348455879e-06,
      "loss": 0.5405,
      "num_input_tokens_seen": 19641344,
      "step": 30035
    },
    {
      "epoch": 15.744234800838575,
      "grad_norm": 0.4564768373966217,
      "learning_rate": 6.586617418817828e-06,
      "loss": 0.3807,
      "num_input_tokens_seen": 19644320,
      "step": 30040
    },
    {
      "epoch": 15.74685534591195,
      "grad_norm": 0.33803778886795044,
      "learning_rate": 6.578885204946439e-06,
      "loss": 0.3334,
      "num_input_tokens_seen": 19648512,
      "step": 30045
    },
    {
      "epoch": 15.749475890985325,
      "grad_norm": 0.520125687122345,
      "learning_rate": 6.571156844562098e-06,
      "loss": 0.5291,
      "num_input_tokens_seen": 19651296,
      "step": 30050
    },
    {
      "epoch": 15.7520964360587,
      "grad_norm": 0.5856311917304993,
      "learning_rate": 6.563432339281497e-06,
      "loss": 0.4672,
      "num_input_tokens_seen": 19654080,
      "step": 30055
    },
    {
      "epoch": 15.754716981132075,
      "grad_norm": 0.44319263100624084,
      "learning_rate": 6.555711690720517e-06,
      "loss": 0.4664,
      "num_input_tokens_seen": 19658240,
      "step": 30060
    },
    {
      "epoch": 15.75733752620545,
      "grad_norm": 0.5197641253471375,
      "learning_rate": 6.547994900494234e-06,
      "loss": 0.5711,
      "num_input_tokens_seen": 19661792,
      "step": 30065
    },
    {
      "epoch": 15.759958071278826,
      "grad_norm": 1.0414671897888184,
      "learning_rate": 6.540281970216899e-06,
      "loss": 0.4711,
      "num_input_tokens_seen": 19664288,
      "step": 30070
    },
    {
      "epoch": 15.7625786163522,
      "grad_norm": 0.5087805390357971,
      "learning_rate": 6.5325729015019845e-06,
      "loss": 0.4737,
      "num_input_tokens_seen": 19670656,
      "step": 30075
    },
    {
      "epoch": 15.765199161425576,
      "grad_norm": 0.40177059173583984,
      "learning_rate": 6.524867695962122e-06,
      "loss": 0.4761,
      "num_input_tokens_seen": 19674048,
      "step": 30080
    },
    {
      "epoch": 15.767819706498951,
      "grad_norm": 0.7763234376907349,
      "learning_rate": 6.517166355209165e-06,
      "loss": 0.3663,
      "num_input_tokens_seen": 19676896,
      "step": 30085
    },
    {
      "epoch": 15.770440251572326,
      "grad_norm": 0.2715233862400055,
      "learning_rate": 6.50946888085415e-06,
      "loss": 0.4155,
      "num_input_tokens_seen": 19679392,
      "step": 30090
    },
    {
      "epoch": 15.773060796645701,
      "grad_norm": 0.5107837319374084,
      "learning_rate": 6.5017752745072855e-06,
      "loss": 0.4652,
      "num_input_tokens_seen": 19682560,
      "step": 30095
    },
    {
      "epoch": 15.775681341719078,
      "grad_norm": 0.6716918349266052,
      "learning_rate": 6.494085537777994e-06,
      "loss": 0.4405,
      "num_input_tokens_seen": 19685312,
      "step": 30100
    },
    {
      "epoch": 15.778301886792454,
      "grad_norm": 1.1022043228149414,
      "learning_rate": 6.48639967227489e-06,
      "loss": 0.4116,
      "num_input_tokens_seen": 19688064,
      "step": 30105
    },
    {
      "epoch": 15.780922431865829,
      "grad_norm": 0.39793631434440613,
      "learning_rate": 6.47871767960575e-06,
      "loss": 0.4659,
      "num_input_tokens_seen": 19690976,
      "step": 30110
    },
    {
      "epoch": 15.783542976939204,
      "grad_norm": 0.46136876940727234,
      "learning_rate": 6.471039561377581e-06,
      "loss": 0.4849,
      "num_input_tokens_seen": 19693728,
      "step": 30115
    },
    {
      "epoch": 15.786163522012579,
      "grad_norm": 0.704239010810852,
      "learning_rate": 6.463365319196538e-06,
      "loss": 0.4927,
      "num_input_tokens_seen": 19700864,
      "step": 30120
    },
    {
      "epoch": 15.788784067085954,
      "grad_norm": 0.5735439658164978,
      "learning_rate": 6.4556949546679905e-06,
      "loss": 0.356,
      "num_input_tokens_seen": 19703488,
      "step": 30125
    },
    {
      "epoch": 15.79140461215933,
      "grad_norm": 0.3931128680706024,
      "learning_rate": 6.448028469396497e-06,
      "loss": 0.3578,
      "num_input_tokens_seen": 19706944,
      "step": 30130
    },
    {
      "epoch": 15.794025157232705,
      "grad_norm": 0.32343414425849915,
      "learning_rate": 6.440365864985801e-06,
      "loss": 0.4166,
      "num_input_tokens_seen": 19710400,
      "step": 30135
    },
    {
      "epoch": 15.79664570230608,
      "grad_norm": 0.5545079112052917,
      "learning_rate": 6.432707143038818e-06,
      "loss": 0.427,
      "num_input_tokens_seen": 19713056,
      "step": 30140
    },
    {
      "epoch": 15.799266247379455,
      "grad_norm": 0.44789063930511475,
      "learning_rate": 6.4250523051576834e-06,
      "loss": 0.4562,
      "num_input_tokens_seen": 19717248,
      "step": 30145
    },
    {
      "epoch": 15.80188679245283,
      "grad_norm": 0.8703071475028992,
      "learning_rate": 6.417401352943686e-06,
      "loss": 0.4696,
      "num_input_tokens_seen": 19719552,
      "step": 30150
    },
    {
      "epoch": 15.804507337526205,
      "grad_norm": 0.5948378443717957,
      "learning_rate": 6.409754287997322e-06,
      "loss": 0.4927,
      "num_input_tokens_seen": 19722304,
      "step": 30155
    },
    {
      "epoch": 15.80712788259958,
      "grad_norm": 0.5162750482559204,
      "learning_rate": 6.40211111191828e-06,
      "loss": 0.4513,
      "num_input_tokens_seen": 19725440,
      "step": 30160
    },
    {
      "epoch": 15.809748427672956,
      "grad_norm": 0.7373701333999634,
      "learning_rate": 6.394471826305409e-06,
      "loss": 0.4575,
      "num_input_tokens_seen": 19728864,
      "step": 30165
    },
    {
      "epoch": 15.81236897274633,
      "grad_norm": 0.6603497266769409,
      "learning_rate": 6.3868364327567795e-06,
      "loss": 0.4344,
      "num_input_tokens_seen": 19732192,
      "step": 30170
    },
    {
      "epoch": 15.814989517819706,
      "grad_norm": 0.35878223180770874,
      "learning_rate": 6.379204932869606e-06,
      "loss": 0.448,
      "num_input_tokens_seen": 19735040,
      "step": 30175
    },
    {
      "epoch": 15.817610062893081,
      "grad_norm": 0.7881470322608948,
      "learning_rate": 6.371577328240327e-06,
      "loss": 0.4217,
      "num_input_tokens_seen": 19738368,
      "step": 30180
    },
    {
      "epoch": 15.820230607966456,
      "grad_norm": 0.6355002522468567,
      "learning_rate": 6.363953620464547e-06,
      "loss": 0.3533,
      "num_input_tokens_seen": 19742176,
      "step": 30185
    },
    {
      "epoch": 15.822851153039831,
      "grad_norm": 0.3912878930568695,
      "learning_rate": 6.356333811137064e-06,
      "loss": 0.4038,
      "num_input_tokens_seen": 19745760,
      "step": 30190
    },
    {
      "epoch": 15.825471698113208,
      "grad_norm": 0.4002615511417389,
      "learning_rate": 6.34871790185185e-06,
      "loss": 0.3963,
      "num_input_tokens_seen": 19748832,
      "step": 30195
    },
    {
      "epoch": 15.828092243186584,
      "grad_norm": 0.5267706513404846,
      "learning_rate": 6.341105894202057e-06,
      "loss": 0.4378,
      "num_input_tokens_seen": 19752000,
      "step": 30200
    },
    {
      "epoch": 15.830712788259959,
      "grad_norm": 0.3549883961677551,
      "learning_rate": 6.333497789780041e-06,
      "loss": 0.4685,
      "num_input_tokens_seen": 19755744,
      "step": 30205
    },
    {
      "epoch": 15.833333333333334,
      "grad_norm": 0.43710917234420776,
      "learning_rate": 6.325893590177329e-06,
      "loss": 0.4847,
      "num_input_tokens_seen": 19759072,
      "step": 30210
    },
    {
      "epoch": 15.835953878406709,
      "grad_norm": 0.36233681440353394,
      "learning_rate": 6.318293296984631e-06,
      "loss": 0.5049,
      "num_input_tokens_seen": 19762432,
      "step": 30215
    },
    {
      "epoch": 15.838574423480084,
      "grad_norm": 0.5158885717391968,
      "learning_rate": 6.3106969117918495e-06,
      "loss": 0.4203,
      "num_input_tokens_seen": 19765248,
      "step": 30220
    },
    {
      "epoch": 15.84119496855346,
      "grad_norm": 0.3842511773109436,
      "learning_rate": 6.303104436188057e-06,
      "loss": 0.4167,
      "num_input_tokens_seen": 19768224,
      "step": 30225
    },
    {
      "epoch": 15.843815513626835,
      "grad_norm": 0.4345457851886749,
      "learning_rate": 6.2955158717615036e-06,
      "loss": 0.4238,
      "num_input_tokens_seen": 19771232,
      "step": 30230
    },
    {
      "epoch": 15.84643605870021,
      "grad_norm": 0.46701985597610474,
      "learning_rate": 6.287931220099638e-06,
      "loss": 0.4245,
      "num_input_tokens_seen": 19774016,
      "step": 30235
    },
    {
      "epoch": 15.849056603773585,
      "grad_norm": 0.4628138840198517,
      "learning_rate": 6.280350482789082e-06,
      "loss": 0.5877,
      "num_input_tokens_seen": 19777568,
      "step": 30240
    },
    {
      "epoch": 15.85167714884696,
      "grad_norm": 1.0239859819412231,
      "learning_rate": 6.272773661415645e-06,
      "loss": 0.3659,
      "num_input_tokens_seen": 19780448,
      "step": 30245
    },
    {
      "epoch": 15.854297693920335,
      "grad_norm": 0.5768481492996216,
      "learning_rate": 6.26520075756431e-06,
      "loss": 0.3763,
      "num_input_tokens_seen": 19783808,
      "step": 30250
    },
    {
      "epoch": 15.85691823899371,
      "grad_norm": 0.3456646800041199,
      "learning_rate": 6.2576317728192304e-06,
      "loss": 0.3886,
      "num_input_tokens_seen": 19787136,
      "step": 30255
    },
    {
      "epoch": 15.859538784067086,
      "grad_norm": 0.6198771595954895,
      "learning_rate": 6.250066708763761e-06,
      "loss": 0.4869,
      "num_input_tokens_seen": 19789920,
      "step": 30260
    },
    {
      "epoch": 15.86215932914046,
      "grad_norm": 0.3061290681362152,
      "learning_rate": 6.242505566980422e-06,
      "loss": 0.4616,
      "num_input_tokens_seen": 19793024,
      "step": 30265
    },
    {
      "epoch": 15.864779874213836,
      "grad_norm": 0.866246223449707,
      "learning_rate": 6.234948349050931e-06,
      "loss": 0.431,
      "num_input_tokens_seen": 19796672,
      "step": 30270
    },
    {
      "epoch": 15.867400419287211,
      "grad_norm": 0.5134739279747009,
      "learning_rate": 6.227395056556162e-06,
      "loss": 0.3922,
      "num_input_tokens_seen": 19798944,
      "step": 30275
    },
    {
      "epoch": 15.870020964360586,
      "grad_norm": 4.118712902069092,
      "learning_rate": 6.219845691076173e-06,
      "loss": 0.4769,
      "num_input_tokens_seen": 19801664,
      "step": 30280
    },
    {
      "epoch": 15.872641509433961,
      "grad_norm": 0.41451576352119446,
      "learning_rate": 6.212300254190206e-06,
      "loss": 0.5201,
      "num_input_tokens_seen": 19804384,
      "step": 30285
    },
    {
      "epoch": 15.875262054507338,
      "grad_norm": 0.44892823696136475,
      "learning_rate": 6.204758747476688e-06,
      "loss": 0.4868,
      "num_input_tokens_seen": 19807744,
      "step": 30290
    },
    {
      "epoch": 15.877882599580714,
      "grad_norm": 0.49456652998924255,
      "learning_rate": 6.1972211725132095e-06,
      "loss": 0.291,
      "num_input_tokens_seen": 19810336,
      "step": 30295
    },
    {
      "epoch": 15.880503144654089,
      "grad_norm": 0.3132758140563965,
      "learning_rate": 6.189687530876559e-06,
      "loss": 0.4006,
      "num_input_tokens_seen": 19813440,
      "step": 30300
    },
    {
      "epoch": 15.883123689727464,
      "grad_norm": 0.5590085387229919,
      "learning_rate": 6.182157824142676e-06,
      "loss": 0.5733,
      "num_input_tokens_seen": 19816128,
      "step": 30305
    },
    {
      "epoch": 15.885744234800839,
      "grad_norm": 0.6367961168289185,
      "learning_rate": 6.174632053886681e-06,
      "loss": 0.3474,
      "num_input_tokens_seen": 19818656,
      "step": 30310
    },
    {
      "epoch": 15.888364779874214,
      "grad_norm": 0.3573092818260193,
      "learning_rate": 6.167110221682893e-06,
      "loss": 0.4506,
      "num_input_tokens_seen": 19821856,
      "step": 30315
    },
    {
      "epoch": 15.89098532494759,
      "grad_norm": 0.6286327242851257,
      "learning_rate": 6.159592329104788e-06,
      "loss": 0.4656,
      "num_input_tokens_seen": 19824864,
      "step": 30320
    },
    {
      "epoch": 15.893605870020965,
      "grad_norm": 0.6583585739135742,
      "learning_rate": 6.152078377725032e-06,
      "loss": 0.3657,
      "num_input_tokens_seen": 19828352,
      "step": 30325
    },
    {
      "epoch": 15.89622641509434,
      "grad_norm": 0.5470496416091919,
      "learning_rate": 6.144568369115453e-06,
      "loss": 0.4818,
      "num_input_tokens_seen": 19830848,
      "step": 30330
    },
    {
      "epoch": 15.898846960167715,
      "grad_norm": 0.37651094794273376,
      "learning_rate": 6.137062304847046e-06,
      "loss": 0.4582,
      "num_input_tokens_seen": 19836448,
      "step": 30335
    },
    {
      "epoch": 15.90146750524109,
      "grad_norm": 0.5627879500389099,
      "learning_rate": 6.129560186490008e-06,
      "loss": 0.3289,
      "num_input_tokens_seen": 19839296,
      "step": 30340
    },
    {
      "epoch": 15.904088050314465,
      "grad_norm": 0.5496777296066284,
      "learning_rate": 6.122062015613694e-06,
      "loss": 0.4075,
      "num_input_tokens_seen": 19842176,
      "step": 30345
    },
    {
      "epoch": 15.90670859538784,
      "grad_norm": 0.5223483443260193,
      "learning_rate": 6.114567793786641e-06,
      "loss": 0.3384,
      "num_input_tokens_seen": 19845216,
      "step": 30350
    },
    {
      "epoch": 15.909329140461216,
      "grad_norm": 0.2962462306022644,
      "learning_rate": 6.107077522576543e-06,
      "loss": 0.5204,
      "num_input_tokens_seen": 19848064,
      "step": 30355
    },
    {
      "epoch": 15.91194968553459,
      "grad_norm": 0.5542677044868469,
      "learning_rate": 6.099591203550292e-06,
      "loss": 0.4279,
      "num_input_tokens_seen": 19850912,
      "step": 30360
    },
    {
      "epoch": 15.914570230607966,
      "grad_norm": 0.5332052111625671,
      "learning_rate": 6.092108838273927e-06,
      "loss": 0.4882,
      "num_input_tokens_seen": 19854688,
      "step": 30365
    },
    {
      "epoch": 15.917190775681341,
      "grad_norm": 0.5892075300216675,
      "learning_rate": 6.08463042831268e-06,
      "loss": 0.4529,
      "num_input_tokens_seen": 19857792,
      "step": 30370
    },
    {
      "epoch": 15.919811320754716,
      "grad_norm": 0.4466644823551178,
      "learning_rate": 6.0771559752309496e-06,
      "loss": 0.4465,
      "num_input_tokens_seen": 19860672,
      "step": 30375
    },
    {
      "epoch": 15.922431865828091,
      "grad_norm": 0.43755555152893066,
      "learning_rate": 6.069685480592313e-06,
      "loss": 0.4233,
      "num_input_tokens_seen": 19863584,
      "step": 30380
    },
    {
      "epoch": 15.925052410901468,
      "grad_norm": 0.6701103448867798,
      "learning_rate": 6.062218945959497e-06,
      "loss": 0.485,
      "num_input_tokens_seen": 19866912,
      "step": 30385
    },
    {
      "epoch": 15.927672955974844,
      "grad_norm": 0.3960433900356293,
      "learning_rate": 6.054756372894435e-06,
      "loss": 0.6199,
      "num_input_tokens_seen": 19869920,
      "step": 30390
    },
    {
      "epoch": 15.930293501048219,
      "grad_norm": 0.5333653688430786,
      "learning_rate": 6.04729776295819e-06,
      "loss": 0.6027,
      "num_input_tokens_seen": 19872736,
      "step": 30395
    },
    {
      "epoch": 15.932914046121594,
      "grad_norm": 0.6490904092788696,
      "learning_rate": 6.0398431177110306e-06,
      "loss": 0.5445,
      "num_input_tokens_seen": 19875584,
      "step": 30400
    },
    {
      "epoch": 15.935534591194969,
      "grad_norm": 0.41023388504981995,
      "learning_rate": 6.032392438712389e-06,
      "loss": 0.4742,
      "num_input_tokens_seen": 19879168,
      "step": 30405
    },
    {
      "epoch": 15.938155136268344,
      "grad_norm": 0.662083625793457,
      "learning_rate": 6.024945727520847e-06,
      "loss": 0.5737,
      "num_input_tokens_seen": 19882528,
      "step": 30410
    },
    {
      "epoch": 15.94077568134172,
      "grad_norm": 0.6488873958587646,
      "learning_rate": 6.01750298569419e-06,
      "loss": 0.4029,
      "num_input_tokens_seen": 19885440,
      "step": 30415
    },
    {
      "epoch": 15.943396226415095,
      "grad_norm": 0.790795624256134,
      "learning_rate": 6.010064214789335e-06,
      "loss": 0.3573,
      "num_input_tokens_seen": 19888320,
      "step": 30420
    },
    {
      "epoch": 15.94601677148847,
      "grad_norm": 0.4070033133029938,
      "learning_rate": 6.002629416362399e-06,
      "loss": 0.5646,
      "num_input_tokens_seen": 19891520,
      "step": 30425
    },
    {
      "epoch": 15.948637316561845,
      "grad_norm": 0.7273432612419128,
      "learning_rate": 5.995198591968662e-06,
      "loss": 0.3386,
      "num_input_tokens_seen": 19894688,
      "step": 30430
    },
    {
      "epoch": 15.95125786163522,
      "grad_norm": 0.7916163802146912,
      "learning_rate": 5.987771743162554e-06,
      "loss": 0.4555,
      "num_input_tokens_seen": 19897408,
      "step": 30435
    },
    {
      "epoch": 15.953878406708595,
      "grad_norm": 0.5166029930114746,
      "learning_rate": 5.9803488714976955e-06,
      "loss": 0.5584,
      "num_input_tokens_seen": 19901056,
      "step": 30440
    },
    {
      "epoch": 15.95649895178197,
      "grad_norm": 0.23594677448272705,
      "learning_rate": 5.972929978526872e-06,
      "loss": 0.299,
      "num_input_tokens_seen": 19905056,
      "step": 30445
    },
    {
      "epoch": 15.959119496855346,
      "grad_norm": 0.3938003480434418,
      "learning_rate": 5.965515065802019e-06,
      "loss": 0.4064,
      "num_input_tokens_seen": 19908160,
      "step": 30450
    },
    {
      "epoch": 15.96174004192872,
      "grad_norm": 0.4675661623477936,
      "learning_rate": 5.958104134874254e-06,
      "loss": 0.3861,
      "num_input_tokens_seen": 19911040,
      "step": 30455
    },
    {
      "epoch": 15.964360587002096,
      "grad_norm": 0.40331897139549255,
      "learning_rate": 5.950697187293872e-06,
      "loss": 0.4343,
      "num_input_tokens_seen": 19914720,
      "step": 30460
    },
    {
      "epoch": 15.966981132075471,
      "grad_norm": 0.594250500202179,
      "learning_rate": 5.943294224610305e-06,
      "loss": 0.5056,
      "num_input_tokens_seen": 19917312,
      "step": 30465
    },
    {
      "epoch": 15.969601677148846,
      "grad_norm": 0.6069393754005432,
      "learning_rate": 5.935895248372175e-06,
      "loss": 0.5361,
      "num_input_tokens_seen": 19920480,
      "step": 30470
    },
    {
      "epoch": 15.972222222222221,
      "grad_norm": 0.67607581615448,
      "learning_rate": 5.928500260127273e-06,
      "loss": 0.3884,
      "num_input_tokens_seen": 19923104,
      "step": 30475
    },
    {
      "epoch": 15.974842767295598,
      "grad_norm": 0.3359546661376953,
      "learning_rate": 5.921109261422531e-06,
      "loss": 0.4673,
      "num_input_tokens_seen": 19926688,
      "step": 30480
    },
    {
      "epoch": 15.977463312368974,
      "grad_norm": 0.4146633744239807,
      "learning_rate": 5.913722253804071e-06,
      "loss": 0.3356,
      "num_input_tokens_seen": 19930624,
      "step": 30485
    },
    {
      "epoch": 15.980083857442349,
      "grad_norm": 0.33619534969329834,
      "learning_rate": 5.906339238817163e-06,
      "loss": 0.485,
      "num_input_tokens_seen": 19933984,
      "step": 30490
    },
    {
      "epoch": 15.982704402515724,
      "grad_norm": 0.5549284219741821,
      "learning_rate": 5.8989602180062554e-06,
      "loss": 0.3639,
      "num_input_tokens_seen": 19936288,
      "step": 30495
    },
    {
      "epoch": 15.985324947589099,
      "grad_norm": 0.6226898431777954,
      "learning_rate": 5.891585192914953e-06,
      "loss": 0.4591,
      "num_input_tokens_seen": 19939648,
      "step": 30500
    },
    {
      "epoch": 15.987945492662474,
      "grad_norm": 0.5434470176696777,
      "learning_rate": 5.884214165086036e-06,
      "loss": 0.4651,
      "num_input_tokens_seen": 19942848,
      "step": 30505
    },
    {
      "epoch": 15.99056603773585,
      "grad_norm": 0.4846609830856323,
      "learning_rate": 5.876847136061428e-06,
      "loss": 0.4832,
      "num_input_tokens_seen": 19946176,
      "step": 30510
    },
    {
      "epoch": 15.993186582809225,
      "grad_norm": 0.6718912124633789,
      "learning_rate": 5.869484107382228e-06,
      "loss": 0.3332,
      "num_input_tokens_seen": 19948800,
      "step": 30515
    },
    {
      "epoch": 15.9958071278826,
      "grad_norm": 0.6270756125450134,
      "learning_rate": 5.862125080588696e-06,
      "loss": 0.4367,
      "num_input_tokens_seen": 19952288,
      "step": 30520
    },
    {
      "epoch": 15.998427672955975,
      "grad_norm": 0.6556290984153748,
      "learning_rate": 5.854770057220263e-06,
      "loss": 0.3818,
      "num_input_tokens_seen": 19955424,
      "step": 30525
    },
    {
      "epoch": 16.0,
      "eval_loss": 0.4753433167934418,
      "eval_runtime": 9.8013,
      "eval_samples_per_second": 86.519,
      "eval_steps_per_second": 21.63,
      "num_input_tokens_seen": 19956544,
      "step": 30528
    },
    {
      "epoch": 16.00104821802935,
      "grad_norm": 0.5864355564117432,
      "learning_rate": 5.8474190388155216e-06,
      "loss": 0.4582,
      "num_input_tokens_seen": 19957376,
      "step": 30530
    },
    {
      "epoch": 16.003668763102727,
      "grad_norm": 0.37382248044013977,
      "learning_rate": 5.840072026912205e-06,
      "loss": 0.5254,
      "num_input_tokens_seen": 19960768,
      "step": 30535
    },
    {
      "epoch": 16.0062893081761,
      "grad_norm": 0.5126827359199524,
      "learning_rate": 5.83272902304724e-06,
      "loss": 0.3487,
      "num_input_tokens_seen": 19964416,
      "step": 30540
    },
    {
      "epoch": 16.008909853249477,
      "grad_norm": 0.498019814491272,
      "learning_rate": 5.825390028756683e-06,
      "loss": 0.4787,
      "num_input_tokens_seen": 19967904,
      "step": 30545
    },
    {
      "epoch": 16.01153039832285,
      "grad_norm": 0.6431699395179749,
      "learning_rate": 5.818055045575776e-06,
      "loss": 0.6109,
      "num_input_tokens_seen": 19971872,
      "step": 30550
    },
    {
      "epoch": 16.014150943396228,
      "grad_norm": 0.5021823644638062,
      "learning_rate": 5.810724075038917e-06,
      "loss": 0.3606,
      "num_input_tokens_seen": 19975072,
      "step": 30555
    },
    {
      "epoch": 16.0167714884696,
      "grad_norm": 1.6609052419662476,
      "learning_rate": 5.803397118679665e-06,
      "loss": 0.588,
      "num_input_tokens_seen": 19978144,
      "step": 30560
    },
    {
      "epoch": 16.019392033542978,
      "grad_norm": 0.40726587176322937,
      "learning_rate": 5.796074178030727e-06,
      "loss": 0.3553,
      "num_input_tokens_seen": 19981216,
      "step": 30565
    },
    {
      "epoch": 16.02201257861635,
      "grad_norm": 0.5354470610618591,
      "learning_rate": 5.788755254623976e-06,
      "loss": 0.5913,
      "num_input_tokens_seen": 19984896,
      "step": 30570
    },
    {
      "epoch": 16.02463312368973,
      "grad_norm": 0.3602522611618042,
      "learning_rate": 5.7814403499904525e-06,
      "loss": 0.4105,
      "num_input_tokens_seen": 19988256,
      "step": 30575
    },
    {
      "epoch": 16.0272536687631,
      "grad_norm": 0.8309329152107239,
      "learning_rate": 5.774129465660352e-06,
      "loss": 0.4608,
      "num_input_tokens_seen": 19991744,
      "step": 30580
    },
    {
      "epoch": 16.02987421383648,
      "grad_norm": 0.4274003207683563,
      "learning_rate": 5.7668226031630246e-06,
      "loss": 0.4104,
      "num_input_tokens_seen": 19995104,
      "step": 30585
    },
    {
      "epoch": 16.032494758909852,
      "grad_norm": 0.5352482795715332,
      "learning_rate": 5.759519764026999e-06,
      "loss": 0.4249,
      "num_input_tokens_seen": 19997472,
      "step": 30590
    },
    {
      "epoch": 16.03511530398323,
      "grad_norm": 0.8553908467292786,
      "learning_rate": 5.7522209497799185e-06,
      "loss": 0.4365,
      "num_input_tokens_seen": 19999776,
      "step": 30595
    },
    {
      "epoch": 16.037735849056602,
      "grad_norm": 0.8283685445785522,
      "learning_rate": 5.7449261619486246e-06,
      "loss": 0.5033,
      "num_input_tokens_seen": 20002848,
      "step": 30600
    },
    {
      "epoch": 16.04035639412998,
      "grad_norm": 0.3747755289077759,
      "learning_rate": 5.737635402059105e-06,
      "loss": 0.3025,
      "num_input_tokens_seen": 20006176,
      "step": 30605
    },
    {
      "epoch": 16.042976939203353,
      "grad_norm": 0.6683638691902161,
      "learning_rate": 5.730348671636501e-06,
      "loss": 0.392,
      "num_input_tokens_seen": 20009088,
      "step": 30610
    },
    {
      "epoch": 16.04559748427673,
      "grad_norm": 0.53716641664505,
      "learning_rate": 5.723065972205124e-06,
      "loss": 0.3541,
      "num_input_tokens_seen": 20012000,
      "step": 30615
    },
    {
      "epoch": 16.048218029350103,
      "grad_norm": 0.4783799648284912,
      "learning_rate": 5.715787305288423e-06,
      "loss": 0.599,
      "num_input_tokens_seen": 20016416,
      "step": 30620
    },
    {
      "epoch": 16.05083857442348,
      "grad_norm": 0.4067259728908539,
      "learning_rate": 5.7085126724090024e-06,
      "loss": 0.461,
      "num_input_tokens_seen": 20020608,
      "step": 30625
    },
    {
      "epoch": 16.053459119496857,
      "grad_norm": 0.5388728380203247,
      "learning_rate": 5.7012420750886445e-06,
      "loss": 0.5479,
      "num_input_tokens_seen": 20023360,
      "step": 30630
    },
    {
      "epoch": 16.05607966457023,
      "grad_norm": 0.2820075452327728,
      "learning_rate": 5.693975514848271e-06,
      "loss": 0.3717,
      "num_input_tokens_seen": 20026240,
      "step": 30635
    },
    {
      "epoch": 16.058700209643607,
      "grad_norm": 0.42541801929473877,
      "learning_rate": 5.686712993207971e-06,
      "loss": 0.526,
      "num_input_tokens_seen": 20029568,
      "step": 30640
    },
    {
      "epoch": 16.06132075471698,
      "grad_norm": 0.5089129209518433,
      "learning_rate": 5.679454511686977e-06,
      "loss": 0.5512,
      "num_input_tokens_seen": 20032288,
      "step": 30645
    },
    {
      "epoch": 16.063941299790358,
      "grad_norm": 0.3929283916950226,
      "learning_rate": 5.672200071803671e-06,
      "loss": 0.3806,
      "num_input_tokens_seen": 20035584,
      "step": 30650
    },
    {
      "epoch": 16.06656184486373,
      "grad_norm": 0.5651609301567078,
      "learning_rate": 5.664949675075607e-06,
      "loss": 0.3536,
      "num_input_tokens_seen": 20038912,
      "step": 30655
    },
    {
      "epoch": 16.069182389937108,
      "grad_norm": 0.6940696239471436,
      "learning_rate": 5.657703323019486e-06,
      "loss": 0.5682,
      "num_input_tokens_seen": 20042656,
      "step": 30660
    },
    {
      "epoch": 16.07180293501048,
      "grad_norm": 0.5142884254455566,
      "learning_rate": 5.65046101715116e-06,
      "loss": 0.5257,
      "num_input_tokens_seen": 20046528,
      "step": 30665
    },
    {
      "epoch": 16.07442348008386,
      "grad_norm": 0.98605877161026,
      "learning_rate": 5.643222758985644e-06,
      "loss": 0.3735,
      "num_input_tokens_seen": 20049600,
      "step": 30670
    },
    {
      "epoch": 16.07704402515723,
      "grad_norm": 0.5102974772453308,
      "learning_rate": 5.635988550037094e-06,
      "loss": 0.4907,
      "num_input_tokens_seen": 20052480,
      "step": 30675
    },
    {
      "epoch": 16.07966457023061,
      "grad_norm": 0.5339467525482178,
      "learning_rate": 5.628758391818817e-06,
      "loss": 0.553,
      "num_input_tokens_seen": 20056128,
      "step": 30680
    },
    {
      "epoch": 16.082285115303982,
      "grad_norm": 0.27048617601394653,
      "learning_rate": 5.621532285843287e-06,
      "loss": 0.3991,
      "num_input_tokens_seen": 20059264,
      "step": 30685
    },
    {
      "epoch": 16.08490566037736,
      "grad_norm": 0.44199177622795105,
      "learning_rate": 5.614310233622119e-06,
      "loss": 0.4299,
      "num_input_tokens_seen": 20062112,
      "step": 30690
    },
    {
      "epoch": 16.087526205450732,
      "grad_norm": 0.4852997064590454,
      "learning_rate": 5.607092236666093e-06,
      "loss": 0.5694,
      "num_input_tokens_seen": 20064896,
      "step": 30695
    },
    {
      "epoch": 16.09014675052411,
      "grad_norm": 0.7853811383247375,
      "learning_rate": 5.599878296485125e-06,
      "loss": 0.4557,
      "num_input_tokens_seen": 20067776,
      "step": 30700
    },
    {
      "epoch": 16.092767295597483,
      "grad_norm": 0.5380034446716309,
      "learning_rate": 5.592668414588284e-06,
      "loss": 0.454,
      "num_input_tokens_seen": 20071072,
      "step": 30705
    },
    {
      "epoch": 16.09538784067086,
      "grad_norm": 0.7302622199058533,
      "learning_rate": 5.585462592483801e-06,
      "loss": 0.4372,
      "num_input_tokens_seen": 20073920,
      "step": 30710
    },
    {
      "epoch": 16.098008385744233,
      "grad_norm": 0.2811085283756256,
      "learning_rate": 5.578260831679047e-06,
      "loss": 0.4998,
      "num_input_tokens_seen": 20077952,
      "step": 30715
    },
    {
      "epoch": 16.10062893081761,
      "grad_norm": 0.45490750670433044,
      "learning_rate": 5.571063133680562e-06,
      "loss": 0.5295,
      "num_input_tokens_seen": 20081472,
      "step": 30720
    },
    {
      "epoch": 16.103249475890987,
      "grad_norm": 0.36308789253234863,
      "learning_rate": 5.563869499994004e-06,
      "loss": 0.3588,
      "num_input_tokens_seen": 20084128,
      "step": 30725
    },
    {
      "epoch": 16.10587002096436,
      "grad_norm": 0.4384567439556122,
      "learning_rate": 5.556679932124215e-06,
      "loss": 0.4984,
      "num_input_tokens_seen": 20087488,
      "step": 30730
    },
    {
      "epoch": 16.108490566037737,
      "grad_norm": 0.7794638276100159,
      "learning_rate": 5.549494431575159e-06,
      "loss": 0.5151,
      "num_input_tokens_seen": 20090656,
      "step": 30735
    },
    {
      "epoch": 16.11111111111111,
      "grad_norm": 0.5519747734069824,
      "learning_rate": 5.542312999849961e-06,
      "loss": 0.3913,
      "num_input_tokens_seen": 20093632,
      "step": 30740
    },
    {
      "epoch": 16.113731656184488,
      "grad_norm": 0.43670037388801575,
      "learning_rate": 5.5351356384509e-06,
      "loss": 0.3388,
      "num_input_tokens_seen": 20097312,
      "step": 30745
    },
    {
      "epoch": 16.11635220125786,
      "grad_norm": 0.45282772183418274,
      "learning_rate": 5.527962348879401e-06,
      "loss": 0.6651,
      "num_input_tokens_seen": 20100832,
      "step": 30750
    },
    {
      "epoch": 16.118972746331238,
      "grad_norm": 0.8004561066627502,
      "learning_rate": 5.520793132636027e-06,
      "loss": 0.5402,
      "num_input_tokens_seen": 20104800,
      "step": 30755
    },
    {
      "epoch": 16.12159329140461,
      "grad_norm": 0.29643353819847107,
      "learning_rate": 5.513627991220502e-06,
      "loss": 0.4015,
      "num_input_tokens_seen": 20108544,
      "step": 30760
    },
    {
      "epoch": 16.12421383647799,
      "grad_norm": 0.3898812234401703,
      "learning_rate": 5.506466926131684e-06,
      "loss": 0.2667,
      "num_input_tokens_seen": 20111680,
      "step": 30765
    },
    {
      "epoch": 16.12683438155136,
      "grad_norm": 0.4544254541397095,
      "learning_rate": 5.4993099388675916e-06,
      "loss": 0.3355,
      "num_input_tokens_seen": 20115328,
      "step": 30770
    },
    {
      "epoch": 16.12945492662474,
      "grad_norm": 0.32692065834999084,
      "learning_rate": 5.49215703092539e-06,
      "loss": 0.3536,
      "num_input_tokens_seen": 20118752,
      "step": 30775
    },
    {
      "epoch": 16.132075471698112,
      "grad_norm": 0.5274996757507324,
      "learning_rate": 5.4850082038013755e-06,
      "loss": 0.3301,
      "num_input_tokens_seen": 20121504,
      "step": 30780
    },
    {
      "epoch": 16.13469601677149,
      "grad_norm": 0.32796749472618103,
      "learning_rate": 5.4778634589910115e-06,
      "loss": 0.4847,
      "num_input_tokens_seen": 20125440,
      "step": 30785
    },
    {
      "epoch": 16.137316561844862,
      "grad_norm": 0.5775651931762695,
      "learning_rate": 5.4707227979888834e-06,
      "loss": 0.4603,
      "num_input_tokens_seen": 20128288,
      "step": 30790
    },
    {
      "epoch": 16.13993710691824,
      "grad_norm": 0.5020627975463867,
      "learning_rate": 5.463586222288747e-06,
      "loss": 0.4552,
      "num_input_tokens_seen": 20131328,
      "step": 30795
    },
    {
      "epoch": 16.142557651991613,
      "grad_norm": 0.41718441247940063,
      "learning_rate": 5.456453733383493e-06,
      "loss": 0.2556,
      "num_input_tokens_seen": 20135616,
      "step": 30800
    },
    {
      "epoch": 16.14517819706499,
      "grad_norm": 0.5724943280220032,
      "learning_rate": 5.449325332765151e-06,
      "loss": 0.4314,
      "num_input_tokens_seen": 20139040,
      "step": 30805
    },
    {
      "epoch": 16.147798742138363,
      "grad_norm": 0.3998260200023651,
      "learning_rate": 5.4422010219249e-06,
      "loss": 0.4892,
      "num_input_tokens_seen": 20141984,
      "step": 30810
    },
    {
      "epoch": 16.15041928721174,
      "grad_norm": 0.3555866777896881,
      "learning_rate": 5.435080802353077e-06,
      "loss": 0.381,
      "num_input_tokens_seen": 20144448,
      "step": 30815
    },
    {
      "epoch": 16.153039832285117,
      "grad_norm": 0.6540340781211853,
      "learning_rate": 5.4279646755391325e-06,
      "loss": 0.4263,
      "num_input_tokens_seen": 20150048,
      "step": 30820
    },
    {
      "epoch": 16.15566037735849,
      "grad_norm": 0.600714385509491,
      "learning_rate": 5.420852642971691e-06,
      "loss": 0.386,
      "num_input_tokens_seen": 20153600,
      "step": 30825
    },
    {
      "epoch": 16.158280922431867,
      "grad_norm": 0.5497214198112488,
      "learning_rate": 5.41374470613851e-06,
      "loss": 0.3762,
      "num_input_tokens_seen": 20156448,
      "step": 30830
    },
    {
      "epoch": 16.16090146750524,
      "grad_norm": 0.9021917581558228,
      "learning_rate": 5.40664086652648e-06,
      "loss": 0.4306,
      "num_input_tokens_seen": 20160256,
      "step": 30835
    },
    {
      "epoch": 16.163522012578618,
      "grad_norm": 0.8704761266708374,
      "learning_rate": 5.399541125621646e-06,
      "loss": 0.4776,
      "num_input_tokens_seen": 20163168,
      "step": 30840
    },
    {
      "epoch": 16.16614255765199,
      "grad_norm": 0.36697620153427124,
      "learning_rate": 5.392445484909203e-06,
      "loss": 0.3411,
      "num_input_tokens_seen": 20166048,
      "step": 30845
    },
    {
      "epoch": 16.168763102725368,
      "grad_norm": 0.3814367651939392,
      "learning_rate": 5.385353945873464e-06,
      "loss": 0.4156,
      "num_input_tokens_seen": 20169632,
      "step": 30850
    },
    {
      "epoch": 16.17138364779874,
      "grad_norm": 0.4565071761608124,
      "learning_rate": 5.378266509997909e-06,
      "loss": 0.4656,
      "num_input_tokens_seen": 20173920,
      "step": 30855
    },
    {
      "epoch": 16.17400419287212,
      "grad_norm": 0.650883138179779,
      "learning_rate": 5.371183178765135e-06,
      "loss": 0.5567,
      "num_input_tokens_seen": 20176448,
      "step": 30860
    },
    {
      "epoch": 16.17662473794549,
      "grad_norm": 0.5088510513305664,
      "learning_rate": 5.364103953656904e-06,
      "loss": 0.479,
      "num_input_tokens_seen": 20180448,
      "step": 30865
    },
    {
      "epoch": 16.17924528301887,
      "grad_norm": 0.7589877843856812,
      "learning_rate": 5.35702883615411e-06,
      "loss": 0.4578,
      "num_input_tokens_seen": 20183232,
      "step": 30870
    },
    {
      "epoch": 16.181865828092242,
      "grad_norm": 0.6584289073944092,
      "learning_rate": 5.34995782773679e-06,
      "loss": 0.4352,
      "num_input_tokens_seen": 20186240,
      "step": 30875
    },
    {
      "epoch": 16.18448637316562,
      "grad_norm": 0.357000470161438,
      "learning_rate": 5.3428909298841115e-06,
      "loss": 0.3621,
      "num_input_tokens_seen": 20189888,
      "step": 30880
    },
    {
      "epoch": 16.187106918238992,
      "grad_norm": 0.31775566935539246,
      "learning_rate": 5.335828144074384e-06,
      "loss": 0.4468,
      "num_input_tokens_seen": 20193440,
      "step": 30885
    },
    {
      "epoch": 16.18972746331237,
      "grad_norm": 0.3669510781764984,
      "learning_rate": 5.3287694717850695e-06,
      "loss": 0.4248,
      "num_input_tokens_seen": 20197088,
      "step": 30890
    },
    {
      "epoch": 16.192348008385743,
      "grad_norm": 0.7371019124984741,
      "learning_rate": 5.321714914492759e-06,
      "loss": 0.4575,
      "num_input_tokens_seen": 20200736,
      "step": 30895
    },
    {
      "epoch": 16.19496855345912,
      "grad_norm": 0.5232452750205994,
      "learning_rate": 5.314664473673195e-06,
      "loss": 0.4099,
      "num_input_tokens_seen": 20203712,
      "step": 30900
    },
    {
      "epoch": 16.197589098532493,
      "grad_norm": 0.4723598062992096,
      "learning_rate": 5.307618150801239e-06,
      "loss": 0.4729,
      "num_input_tokens_seen": 20206688,
      "step": 30905
    },
    {
      "epoch": 16.20020964360587,
      "grad_norm": 0.5587232112884521,
      "learning_rate": 5.300575947350897e-06,
      "loss": 0.4766,
      "num_input_tokens_seen": 20210112,
      "step": 30910
    },
    {
      "epoch": 16.202830188679247,
      "grad_norm": 0.40534642338752747,
      "learning_rate": 5.293537864795325e-06,
      "loss": 0.4751,
      "num_input_tokens_seen": 20212768,
      "step": 30915
    },
    {
      "epoch": 16.20545073375262,
      "grad_norm": 0.421760231256485,
      "learning_rate": 5.2865039046068075e-06,
      "loss": 0.5522,
      "num_input_tokens_seen": 20216928,
      "step": 30920
    },
    {
      "epoch": 16.208071278825997,
      "grad_norm": 0.5007307529449463,
      "learning_rate": 5.279474068256767e-06,
      "loss": 0.3728,
      "num_input_tokens_seen": 20219488,
      "step": 30925
    },
    {
      "epoch": 16.21069182389937,
      "grad_norm": 0.6949753165245056,
      "learning_rate": 5.272448357215779e-06,
      "loss": 0.3388,
      "num_input_tokens_seen": 20222592,
      "step": 30930
    },
    {
      "epoch": 16.213312368972748,
      "grad_norm": 0.47497648000717163,
      "learning_rate": 5.26542677295353e-06,
      "loss": 0.577,
      "num_input_tokens_seen": 20227616,
      "step": 30935
    },
    {
      "epoch": 16.21593291404612,
      "grad_norm": 1.1367213726043701,
      "learning_rate": 5.258409316938848e-06,
      "loss": 0.4766,
      "num_input_tokens_seen": 20230240,
      "step": 30940
    },
    {
      "epoch": 16.218553459119498,
      "grad_norm": 0.9489504098892212,
      "learning_rate": 5.251395990639713e-06,
      "loss": 0.4745,
      "num_input_tokens_seen": 20233120,
      "step": 30945
    },
    {
      "epoch": 16.22117400419287,
      "grad_norm": 0.30034318566322327,
      "learning_rate": 5.244386795523235e-06,
      "loss": 0.4828,
      "num_input_tokens_seen": 20236512,
      "step": 30950
    },
    {
      "epoch": 16.22379454926625,
      "grad_norm": 0.3276740610599518,
      "learning_rate": 5.23738173305566e-06,
      "loss": 0.3386,
      "num_input_tokens_seen": 20241024,
      "step": 30955
    },
    {
      "epoch": 16.22641509433962,
      "grad_norm": 0.3658983111381531,
      "learning_rate": 5.230380804702362e-06,
      "loss": 0.3416,
      "num_input_tokens_seen": 20244800,
      "step": 30960
    },
    {
      "epoch": 16.229035639413,
      "grad_norm": 0.2720273733139038,
      "learning_rate": 5.223384011927851e-06,
      "loss": 0.4744,
      "num_input_tokens_seen": 20249152,
      "step": 30965
    },
    {
      "epoch": 16.231656184486372,
      "grad_norm": 0.772251546382904,
      "learning_rate": 5.216391356195779e-06,
      "loss": 0.4982,
      "num_input_tokens_seen": 20252512,
      "step": 30970
    },
    {
      "epoch": 16.23427672955975,
      "grad_norm": 0.5942118167877197,
      "learning_rate": 5.209402838968935e-06,
      "loss": 0.3835,
      "num_input_tokens_seen": 20256224,
      "step": 30975
    },
    {
      "epoch": 16.236897274633122,
      "grad_norm": 0.520510196685791,
      "learning_rate": 5.202418461709233e-06,
      "loss": 0.5468,
      "num_input_tokens_seen": 20258656,
      "step": 30980
    },
    {
      "epoch": 16.2395178197065,
      "grad_norm": 0.3501090109348297,
      "learning_rate": 5.195438225877738e-06,
      "loss": 0.3399,
      "num_input_tokens_seen": 20262528,
      "step": 30985
    },
    {
      "epoch": 16.242138364779873,
      "grad_norm": 0.43781134486198425,
      "learning_rate": 5.18846213293461e-06,
      "loss": 0.4508,
      "num_input_tokens_seen": 20266176,
      "step": 30990
    },
    {
      "epoch": 16.24475890985325,
      "grad_norm": 0.7414273619651794,
      "learning_rate": 5.181490184339185e-06,
      "loss": 0.3451,
      "num_input_tokens_seen": 20269088,
      "step": 30995
    },
    {
      "epoch": 16.247379454926623,
      "grad_norm": 0.8148085474967957,
      "learning_rate": 5.17452238154991e-06,
      "loss": 0.3723,
      "num_input_tokens_seen": 20272384,
      "step": 31000
    },
    {
      "epoch": 16.25,
      "grad_norm": 0.3351942300796509,
      "learning_rate": 5.1675587260243715e-06,
      "loss": 0.4057,
      "num_input_tokens_seen": 20276416,
      "step": 31005
    },
    {
      "epoch": 16.252620545073377,
      "grad_norm": 0.6396675109863281,
      "learning_rate": 5.160599219219295e-06,
      "loss": 0.4177,
      "num_input_tokens_seen": 20279904,
      "step": 31010
    },
    {
      "epoch": 16.25524109014675,
      "grad_norm": 0.3743859827518463,
      "learning_rate": 5.1536438625905185e-06,
      "loss": 0.4108,
      "num_input_tokens_seen": 20283200,
      "step": 31015
    },
    {
      "epoch": 16.257861635220127,
      "grad_norm": 0.7942771911621094,
      "learning_rate": 5.146692657593025e-06,
      "loss": 0.3828,
      "num_input_tokens_seen": 20286336,
      "step": 31020
    },
    {
      "epoch": 16.2604821802935,
      "grad_norm": 0.6480501890182495,
      "learning_rate": 5.139745605680926e-06,
      "loss": 0.4377,
      "num_input_tokens_seen": 20289024,
      "step": 31025
    },
    {
      "epoch": 16.263102725366878,
      "grad_norm": 0.5237941145896912,
      "learning_rate": 5.132802708307469e-06,
      "loss": 0.5325,
      "num_input_tokens_seen": 20291968,
      "step": 31030
    },
    {
      "epoch": 16.26572327044025,
      "grad_norm": 0.5593055486679077,
      "learning_rate": 5.125863966925035e-06,
      "loss": 0.4189,
      "num_input_tokens_seen": 20295072,
      "step": 31035
    },
    {
      "epoch": 16.268343815513628,
      "grad_norm": 0.42080265283584595,
      "learning_rate": 5.118929382985116e-06,
      "loss": 0.3343,
      "num_input_tokens_seen": 20298112,
      "step": 31040
    },
    {
      "epoch": 16.270964360587,
      "grad_norm": 0.36635613441467285,
      "learning_rate": 5.111998957938363e-06,
      "loss": 0.5131,
      "num_input_tokens_seen": 20301664,
      "step": 31045
    },
    {
      "epoch": 16.27358490566038,
      "grad_norm": 0.4729783535003662,
      "learning_rate": 5.105072693234525e-06,
      "loss": 0.4406,
      "num_input_tokens_seen": 20304608,
      "step": 31050
    },
    {
      "epoch": 16.27620545073375,
      "grad_norm": 0.684174120426178,
      "learning_rate": 5.098150590322506e-06,
      "loss": 0.4591,
      "num_input_tokens_seen": 20307840,
      "step": 31055
    },
    {
      "epoch": 16.27882599580713,
      "grad_norm": 0.4832439720630646,
      "learning_rate": 5.091232650650335e-06,
      "loss": 0.3641,
      "num_input_tokens_seen": 20310784,
      "step": 31060
    },
    {
      "epoch": 16.281446540880502,
      "grad_norm": 0.368592232465744,
      "learning_rate": 5.0843188756651645e-06,
      "loss": 0.2768,
      "num_input_tokens_seen": 20313344,
      "step": 31065
    },
    {
      "epoch": 16.28406708595388,
      "grad_norm": 0.7815110683441162,
      "learning_rate": 5.077409266813276e-06,
      "loss": 0.3822,
      "num_input_tokens_seen": 20316256,
      "step": 31070
    },
    {
      "epoch": 16.286687631027252,
      "grad_norm": 0.395510196685791,
      "learning_rate": 5.0705038255400765e-06,
      "loss": 0.4117,
      "num_input_tokens_seen": 20320672,
      "step": 31075
    },
    {
      "epoch": 16.28930817610063,
      "grad_norm": 0.23204196989536285,
      "learning_rate": 5.063602553290106e-06,
      "loss": 0.4325,
      "num_input_tokens_seen": 20324128,
      "step": 31080
    },
    {
      "epoch": 16.291928721174003,
      "grad_norm": 0.5899987816810608,
      "learning_rate": 5.056705451507035e-06,
      "loss": 0.4332,
      "num_input_tokens_seen": 20331520,
      "step": 31085
    },
    {
      "epoch": 16.29454926624738,
      "grad_norm": 0.45303767919540405,
      "learning_rate": 5.049812521633668e-06,
      "loss": 0.3917,
      "num_input_tokens_seen": 20334560,
      "step": 31090
    },
    {
      "epoch": 16.297169811320753,
      "grad_norm": 0.4530288279056549,
      "learning_rate": 5.042923765111907e-06,
      "loss": 0.4632,
      "num_input_tokens_seen": 20337856,
      "step": 31095
    },
    {
      "epoch": 16.29979035639413,
      "grad_norm": 0.4385457932949066,
      "learning_rate": 5.036039183382821e-06,
      "loss": 0.3357,
      "num_input_tokens_seen": 20340224,
      "step": 31100
    },
    {
      "epoch": 16.302410901467507,
      "grad_norm": 0.5002585649490356,
      "learning_rate": 5.02915877788657e-06,
      "loss": 0.456,
      "num_input_tokens_seen": 20343968,
      "step": 31105
    },
    {
      "epoch": 16.30503144654088,
      "grad_norm": 0.32592257857322693,
      "learning_rate": 5.0222825500624625e-06,
      "loss": 0.3995,
      "num_input_tokens_seen": 20348096,
      "step": 31110
    },
    {
      "epoch": 16.307651991614257,
      "grad_norm": 0.596523106098175,
      "learning_rate": 5.015410501348935e-06,
      "loss": 0.5083,
      "num_input_tokens_seen": 20351072,
      "step": 31115
    },
    {
      "epoch": 16.31027253668763,
      "grad_norm": 0.3844846785068512,
      "learning_rate": 5.008542633183527e-06,
      "loss": 0.4252,
      "num_input_tokens_seen": 20354528,
      "step": 31120
    },
    {
      "epoch": 16.312893081761008,
      "grad_norm": 0.4990607500076294,
      "learning_rate": 5.001678947002927e-06,
      "loss": 0.4192,
      "num_input_tokens_seen": 20357664,
      "step": 31125
    },
    {
      "epoch": 16.31551362683438,
      "grad_norm": 0.666278600692749,
      "learning_rate": 4.994819444242943e-06,
      "loss": 0.4015,
      "num_input_tokens_seen": 20361056,
      "step": 31130
    },
    {
      "epoch": 16.318134171907758,
      "grad_norm": 0.5297838449478149,
      "learning_rate": 4.9879641263384956e-06,
      "loss": 0.3362,
      "num_input_tokens_seen": 20364160,
      "step": 31135
    },
    {
      "epoch": 16.32075471698113,
      "grad_norm": 0.5319278836250305,
      "learning_rate": 4.981112994723644e-06,
      "loss": 0.5323,
      "num_input_tokens_seen": 20367072,
      "step": 31140
    },
    {
      "epoch": 16.32337526205451,
      "grad_norm": 0.5452246069908142,
      "learning_rate": 4.974266050831572e-06,
      "loss": 0.3959,
      "num_input_tokens_seen": 20369600,
      "step": 31145
    },
    {
      "epoch": 16.32599580712788,
      "grad_norm": 0.47056522965431213,
      "learning_rate": 4.967423296094567e-06,
      "loss": 0.4274,
      "num_input_tokens_seen": 20372768,
      "step": 31150
    },
    {
      "epoch": 16.32861635220126,
      "grad_norm": 0.5256269574165344,
      "learning_rate": 4.960584731944076e-06,
      "loss": 0.4893,
      "num_input_tokens_seen": 20375680,
      "step": 31155
    },
    {
      "epoch": 16.331236897274632,
      "grad_norm": 0.5261591076850891,
      "learning_rate": 4.953750359810627e-06,
      "loss": 0.4064,
      "num_input_tokens_seen": 20378272,
      "step": 31160
    },
    {
      "epoch": 16.33385744234801,
      "grad_norm": 0.2891083359718323,
      "learning_rate": 4.946920181123904e-06,
      "loss": 0.4102,
      "num_input_tokens_seen": 20381888,
      "step": 31165
    },
    {
      "epoch": 16.336477987421382,
      "grad_norm": 0.32527920603752136,
      "learning_rate": 4.940094197312709e-06,
      "loss": 0.2996,
      "num_input_tokens_seen": 20385856,
      "step": 31170
    },
    {
      "epoch": 16.33909853249476,
      "grad_norm": 0.6673837304115295,
      "learning_rate": 4.933272409804946e-06,
      "loss": 0.4746,
      "num_input_tokens_seen": 20389120,
      "step": 31175
    },
    {
      "epoch": 16.341719077568133,
      "grad_norm": 0.7097439765930176,
      "learning_rate": 4.926454820027663e-06,
      "loss": 0.3992,
      "num_input_tokens_seen": 20392672,
      "step": 31180
    },
    {
      "epoch": 16.34433962264151,
      "grad_norm": 0.5453248620033264,
      "learning_rate": 4.919641429407024e-06,
      "loss": 0.4202,
      "num_input_tokens_seen": 20395232,
      "step": 31185
    },
    {
      "epoch": 16.346960167714883,
      "grad_norm": 0.38868457078933716,
      "learning_rate": 4.912832239368304e-06,
      "loss": 0.3958,
      "num_input_tokens_seen": 20398368,
      "step": 31190
    },
    {
      "epoch": 16.34958071278826,
      "grad_norm": 0.5386355519294739,
      "learning_rate": 4.906027251335918e-06,
      "loss": 0.3675,
      "num_input_tokens_seen": 20400768,
      "step": 31195
    },
    {
      "epoch": 16.352201257861637,
      "grad_norm": 0.34086352586746216,
      "learning_rate": 4.899226466733381e-06,
      "loss": 0.502,
      "num_input_tokens_seen": 20404448,
      "step": 31200
    },
    {
      "epoch": 16.35482180293501,
      "grad_norm": 0.6946951746940613,
      "learning_rate": 4.892429886983346e-06,
      "loss": 0.4995,
      "num_input_tokens_seen": 20407264,
      "step": 31205
    },
    {
      "epoch": 16.357442348008387,
      "grad_norm": 0.5093221664428711,
      "learning_rate": 4.88563751350758e-06,
      "loss": 0.3185,
      "num_input_tokens_seen": 20410272,
      "step": 31210
    },
    {
      "epoch": 16.36006289308176,
      "grad_norm": 0.29358649253845215,
      "learning_rate": 4.878849347726977e-06,
      "loss": 0.4765,
      "num_input_tokens_seen": 20413280,
      "step": 31215
    },
    {
      "epoch": 16.362683438155138,
      "grad_norm": 0.5881283283233643,
      "learning_rate": 4.872065391061526e-06,
      "loss": 0.4857,
      "num_input_tokens_seen": 20415776,
      "step": 31220
    },
    {
      "epoch": 16.36530398322851,
      "grad_norm": 0.6605890393257141,
      "learning_rate": 4.865285644930373e-06,
      "loss": 0.4891,
      "num_input_tokens_seen": 20418944,
      "step": 31225
    },
    {
      "epoch": 16.367924528301888,
      "grad_norm": 0.47656089067459106,
      "learning_rate": 4.858510110751746e-06,
      "loss": 0.5877,
      "num_input_tokens_seen": 20422112,
      "step": 31230
    },
    {
      "epoch": 16.37054507337526,
      "grad_norm": 0.593122124671936,
      "learning_rate": 4.851738789943019e-06,
      "loss": 0.3677,
      "num_input_tokens_seen": 20424736,
      "step": 31235
    },
    {
      "epoch": 16.37316561844864,
      "grad_norm": 0.5357245206832886,
      "learning_rate": 4.844971683920674e-06,
      "loss": 0.4105,
      "num_input_tokens_seen": 20428288,
      "step": 31240
    },
    {
      "epoch": 16.37578616352201,
      "grad_norm": 0.4397966265678406,
      "learning_rate": 4.838208794100316e-06,
      "loss": 0.4697,
      "num_input_tokens_seen": 20432320,
      "step": 31245
    },
    {
      "epoch": 16.37840670859539,
      "grad_norm": 0.8426340818405151,
      "learning_rate": 4.831450121896661e-06,
      "loss": 0.3903,
      "num_input_tokens_seen": 20434976,
      "step": 31250
    },
    {
      "epoch": 16.381027253668762,
      "grad_norm": 0.5449612140655518,
      "learning_rate": 4.824695668723542e-06,
      "loss": 0.4364,
      "num_input_tokens_seen": 20437664,
      "step": 31255
    },
    {
      "epoch": 16.38364779874214,
      "grad_norm": 0.34736505150794983,
      "learning_rate": 4.817945435993915e-06,
      "loss": 0.4411,
      "num_input_tokens_seen": 20440896,
      "step": 31260
    },
    {
      "epoch": 16.386268343815512,
      "grad_norm": 0.3316037654876709,
      "learning_rate": 4.811199425119853e-06,
      "loss": 0.3819,
      "num_input_tokens_seen": 20444064,
      "step": 31265
    },
    {
      "epoch": 16.38888888888889,
      "grad_norm": 0.6101601719856262,
      "learning_rate": 4.804457637512552e-06,
      "loss": 0.4046,
      "num_input_tokens_seen": 20446944,
      "step": 31270
    },
    {
      "epoch": 16.391509433962263,
      "grad_norm": 0.5999465584754944,
      "learning_rate": 4.79772007458231e-06,
      "loss": 0.4535,
      "num_input_tokens_seen": 20449888,
      "step": 31275
    },
    {
      "epoch": 16.39412997903564,
      "grad_norm": 0.41923707723617554,
      "learning_rate": 4.7909867377385424e-06,
      "loss": 0.4072,
      "num_input_tokens_seen": 20453824,
      "step": 31280
    },
    {
      "epoch": 16.396750524109013,
      "grad_norm": 0.5972052216529846,
      "learning_rate": 4.784257628389794e-06,
      "loss": 0.5964,
      "num_input_tokens_seen": 20457184,
      "step": 31285
    },
    {
      "epoch": 16.39937106918239,
      "grad_norm": 0.3652259409427643,
      "learning_rate": 4.7775327479437166e-06,
      "loss": 0.4405,
      "num_input_tokens_seen": 20460864,
      "step": 31290
    },
    {
      "epoch": 16.401991614255767,
      "grad_norm": 0.42116937041282654,
      "learning_rate": 4.770812097807076e-06,
      "loss": 0.3702,
      "num_input_tokens_seen": 20463392,
      "step": 31295
    },
    {
      "epoch": 16.40461215932914,
      "grad_norm": 0.7096717953681946,
      "learning_rate": 4.764095679385766e-06,
      "loss": 0.3946,
      "num_input_tokens_seen": 20466432,
      "step": 31300
    },
    {
      "epoch": 16.407232704402517,
      "grad_norm": 0.3683913052082062,
      "learning_rate": 4.7573834940847745e-06,
      "loss": 0.4862,
      "num_input_tokens_seen": 20469312,
      "step": 31305
    },
    {
      "epoch": 16.40985324947589,
      "grad_norm": 0.46203887462615967,
      "learning_rate": 4.750675543308212e-06,
      "loss": 0.5335,
      "num_input_tokens_seen": 20471936,
      "step": 31310
    },
    {
      "epoch": 16.412473794549268,
      "grad_norm": 0.355192095041275,
      "learning_rate": 4.743971828459307e-06,
      "loss": 0.2877,
      "num_input_tokens_seen": 20474720,
      "step": 31315
    },
    {
      "epoch": 16.41509433962264,
      "grad_norm": 0.44667842984199524,
      "learning_rate": 4.7372723509404025e-06,
      "loss": 0.5393,
      "num_input_tokens_seen": 20478784,
      "step": 31320
    },
    {
      "epoch": 16.417714884696018,
      "grad_norm": 0.5518394112586975,
      "learning_rate": 4.7305771121529615e-06,
      "loss": 0.3543,
      "num_input_tokens_seen": 20481824,
      "step": 31325
    },
    {
      "epoch": 16.42033542976939,
      "grad_norm": 0.5740303993225098,
      "learning_rate": 4.723886113497542e-06,
      "loss": 0.3559,
      "num_input_tokens_seen": 20484704,
      "step": 31330
    },
    {
      "epoch": 16.42295597484277,
      "grad_norm": 0.7077511548995972,
      "learning_rate": 4.7171993563738195e-06,
      "loss": 0.5157,
      "num_input_tokens_seen": 20487936,
      "step": 31335
    },
    {
      "epoch": 16.42557651991614,
      "grad_norm": 0.7152761220932007,
      "learning_rate": 4.7105168421805924e-06,
      "loss": 0.497,
      "num_input_tokens_seen": 20491264,
      "step": 31340
    },
    {
      "epoch": 16.42819706498952,
      "grad_norm": 0.6593232154846191,
      "learning_rate": 4.703838572315767e-06,
      "loss": 0.453,
      "num_input_tokens_seen": 20494112,
      "step": 31345
    },
    {
      "epoch": 16.430817610062892,
      "grad_norm": 0.6329948306083679,
      "learning_rate": 4.697164548176367e-06,
      "loss": 0.4302,
      "num_input_tokens_seen": 20497184,
      "step": 31350
    },
    {
      "epoch": 16.43343815513627,
      "grad_norm": 0.5724102258682251,
      "learning_rate": 4.690494771158519e-06,
      "loss": 0.3662,
      "num_input_tokens_seen": 20500864,
      "step": 31355
    },
    {
      "epoch": 16.436058700209642,
      "grad_norm": 0.5334797501564026,
      "learning_rate": 4.683829242657453e-06,
      "loss": 0.427,
      "num_input_tokens_seen": 20504928,
      "step": 31360
    },
    {
      "epoch": 16.43867924528302,
      "grad_norm": 0.9157803058624268,
      "learning_rate": 4.677167964067533e-06,
      "loss": 0.4796,
      "num_input_tokens_seen": 20507776,
      "step": 31365
    },
    {
      "epoch": 16.441299790356393,
      "grad_norm": 0.5738552808761597,
      "learning_rate": 4.670510936782216e-06,
      "loss": 0.4833,
      "num_input_tokens_seen": 20510944,
      "step": 31370
    },
    {
      "epoch": 16.44392033542977,
      "grad_norm": 0.43393024802207947,
      "learning_rate": 4.6638581621940815e-06,
      "loss": 0.4436,
      "num_input_tokens_seen": 20514784,
      "step": 31375
    },
    {
      "epoch": 16.446540880503143,
      "grad_norm": 0.42606350779533386,
      "learning_rate": 4.657209641694821e-06,
      "loss": 0.3763,
      "num_input_tokens_seen": 20517248,
      "step": 31380
    },
    {
      "epoch": 16.44916142557652,
      "grad_norm": 0.5486792922019958,
      "learning_rate": 4.65056537667522e-06,
      "loss": 0.4308,
      "num_input_tokens_seen": 20520512,
      "step": 31385
    },
    {
      "epoch": 16.451781970649897,
      "grad_norm": 0.40785661339759827,
      "learning_rate": 4.643925368525176e-06,
      "loss": 0.4441,
      "num_input_tokens_seen": 20523328,
      "step": 31390
    },
    {
      "epoch": 16.45440251572327,
      "grad_norm": 0.4414651095867157,
      "learning_rate": 4.637289618633711e-06,
      "loss": 0.3697,
      "num_input_tokens_seen": 20527296,
      "step": 31395
    },
    {
      "epoch": 16.457023060796647,
      "grad_norm": 1.0312579870224,
      "learning_rate": 4.630658128388948e-06,
      "loss": 0.5582,
      "num_input_tokens_seen": 20529952,
      "step": 31400
    },
    {
      "epoch": 16.45964360587002,
      "grad_norm": 0.40181809663772583,
      "learning_rate": 4.624030899178125e-06,
      "loss": 0.4874,
      "num_input_tokens_seen": 20533760,
      "step": 31405
    },
    {
      "epoch": 16.462264150943398,
      "grad_norm": 0.5809487104415894,
      "learning_rate": 4.617407932387569e-06,
      "loss": 0.3801,
      "num_input_tokens_seen": 20536384,
      "step": 31410
    },
    {
      "epoch": 16.46488469601677,
      "grad_norm": 0.33265241980552673,
      "learning_rate": 4.610789229402743e-06,
      "loss": 0.2623,
      "num_input_tokens_seen": 20539744,
      "step": 31415
    },
    {
      "epoch": 16.467505241090148,
      "grad_norm": 0.43117526173591614,
      "learning_rate": 4.604174791608193e-06,
      "loss": 0.3975,
      "num_input_tokens_seen": 20543904,
      "step": 31420
    },
    {
      "epoch": 16.47012578616352,
      "grad_norm": 0.3525616228580475,
      "learning_rate": 4.597564620387587e-06,
      "loss": 0.3295,
      "num_input_tokens_seen": 20547552,
      "step": 31425
    },
    {
      "epoch": 16.4727463312369,
      "grad_norm": 0.42651277780532837,
      "learning_rate": 4.590958717123708e-06,
      "loss": 0.3566,
      "num_input_tokens_seen": 20549984,
      "step": 31430
    },
    {
      "epoch": 16.47536687631027,
      "grad_norm": 0.5311301946640015,
      "learning_rate": 4.5843570831984174e-06,
      "loss": 0.4316,
      "num_input_tokens_seen": 20553280,
      "step": 31435
    },
    {
      "epoch": 16.47798742138365,
      "grad_norm": 0.24395011365413666,
      "learning_rate": 4.577759719992719e-06,
      "loss": 0.4806,
      "num_input_tokens_seen": 20557376,
      "step": 31440
    },
    {
      "epoch": 16.480607966457022,
      "grad_norm": 0.49204757809638977,
      "learning_rate": 4.571166628886692e-06,
      "loss": 0.5226,
      "num_input_tokens_seen": 20560288,
      "step": 31445
    },
    {
      "epoch": 16.4832285115304,
      "grad_norm": 0.2305798977613449,
      "learning_rate": 4.564577811259541e-06,
      "loss": 0.4668,
      "num_input_tokens_seen": 20563392,
      "step": 31450
    },
    {
      "epoch": 16.485849056603772,
      "grad_norm": 0.41776376962661743,
      "learning_rate": 4.557993268489574e-06,
      "loss": 0.4848,
      "num_input_tokens_seen": 20566272,
      "step": 31455
    },
    {
      "epoch": 16.48846960167715,
      "grad_norm": 0.4847241938114166,
      "learning_rate": 4.551413001954205e-06,
      "loss": 0.4309,
      "num_input_tokens_seen": 20568672,
      "step": 31460
    },
    {
      "epoch": 16.491090146750523,
      "grad_norm": 0.5052133798599243,
      "learning_rate": 4.54483701302994e-06,
      "loss": 0.5314,
      "num_input_tokens_seen": 20572096,
      "step": 31465
    },
    {
      "epoch": 16.4937106918239,
      "grad_norm": 0.4185883104801178,
      "learning_rate": 4.538265303092415e-06,
      "loss": 0.3562,
      "num_input_tokens_seen": 20575008,
      "step": 31470
    },
    {
      "epoch": 16.496331236897273,
      "grad_norm": 0.7063867449760437,
      "learning_rate": 4.531697873516338e-06,
      "loss": 0.5762,
      "num_input_tokens_seen": 20578304,
      "step": 31475
    },
    {
      "epoch": 16.49895178197065,
      "grad_norm": 0.5407091379165649,
      "learning_rate": 4.525134725675556e-06,
      "loss": 0.4775,
      "num_input_tokens_seen": 20581248,
      "step": 31480
    },
    {
      "epoch": 16.501572327044027,
      "grad_norm": 0.4756135046482086,
      "learning_rate": 4.518575860943003e-06,
      "loss": 0.3854,
      "num_input_tokens_seen": 20584032,
      "step": 31485
    },
    {
      "epoch": 16.5041928721174,
      "grad_norm": 0.4933280646800995,
      "learning_rate": 4.512021280690712e-06,
      "loss": 0.51,
      "num_input_tokens_seen": 20587872,
      "step": 31490
    },
    {
      "epoch": 16.506813417190777,
      "grad_norm": 0.39662376046180725,
      "learning_rate": 4.505470986289828e-06,
      "loss": 0.4144,
      "num_input_tokens_seen": 20590880,
      "step": 31495
    },
    {
      "epoch": 16.50943396226415,
      "grad_norm": 0.6111515760421753,
      "learning_rate": 4.498924979110605e-06,
      "loss": 0.3153,
      "num_input_tokens_seen": 20593792,
      "step": 31500
    },
    {
      "epoch": 16.512054507337528,
      "grad_norm": 0.29202666878700256,
      "learning_rate": 4.492383260522384e-06,
      "loss": 0.3883,
      "num_input_tokens_seen": 20596992,
      "step": 31505
    },
    {
      "epoch": 16.5146750524109,
      "grad_norm": 0.6009622812271118,
      "learning_rate": 4.485845831893626e-06,
      "loss": 0.4815,
      "num_input_tokens_seen": 20599808,
      "step": 31510
    },
    {
      "epoch": 16.517295597484278,
      "grad_norm": 0.3414047062397003,
      "learning_rate": 4.479312694591878e-06,
      "loss": 0.4534,
      "num_input_tokens_seen": 20603136,
      "step": 31515
    },
    {
      "epoch": 16.51991614255765,
      "grad_norm": 0.721384584903717,
      "learning_rate": 4.4727838499838e-06,
      "loss": 0.3863,
      "num_input_tokens_seen": 20605792,
      "step": 31520
    },
    {
      "epoch": 16.52253668763103,
      "grad_norm": 0.8772330284118652,
      "learning_rate": 4.4662592994351545e-06,
      "loss": 0.5148,
      "num_input_tokens_seen": 20608928,
      "step": 31525
    },
    {
      "epoch": 16.5251572327044,
      "grad_norm": 0.5653486847877502,
      "learning_rate": 4.4597390443108065e-06,
      "loss": 0.4639,
      "num_input_tokens_seen": 20611680,
      "step": 31530
    },
    {
      "epoch": 16.52777777777778,
      "grad_norm": 0.6161595582962036,
      "learning_rate": 4.453223085974712e-06,
      "loss": 0.3329,
      "num_input_tokens_seen": 20614528,
      "step": 31535
    },
    {
      "epoch": 16.530398322851152,
      "grad_norm": 0.49908187985420227,
      "learning_rate": 4.4467114257899425e-06,
      "loss": 0.4337,
      "num_input_tokens_seen": 20617984,
      "step": 31540
    },
    {
      "epoch": 16.53301886792453,
      "grad_norm": 0.4698735773563385,
      "learning_rate": 4.44020406511865e-06,
      "loss": 0.3305,
      "num_input_tokens_seen": 20621952,
      "step": 31545
    },
    {
      "epoch": 16.535639412997902,
      "grad_norm": 1.1988242864608765,
      "learning_rate": 4.43370100532211e-06,
      "loss": 0.5538,
      "num_input_tokens_seen": 20624416,
      "step": 31550
    },
    {
      "epoch": 16.53825995807128,
      "grad_norm": 0.3420881927013397,
      "learning_rate": 4.427202247760695e-06,
      "loss": 0.3495,
      "num_input_tokens_seen": 20627456,
      "step": 31555
    },
    {
      "epoch": 16.540880503144653,
      "grad_norm": 0.526644766330719,
      "learning_rate": 4.420707793793854e-06,
      "loss": 0.4165,
      "num_input_tokens_seen": 20630688,
      "step": 31560
    },
    {
      "epoch": 16.54350104821803,
      "grad_norm": 0.4215386211872101,
      "learning_rate": 4.414217644780172e-06,
      "loss": 0.488,
      "num_input_tokens_seen": 20635552,
      "step": 31565
    },
    {
      "epoch": 16.546121593291403,
      "grad_norm": 2.276231527328491,
      "learning_rate": 4.407731802077294e-06,
      "loss": 0.492,
      "num_input_tokens_seen": 20638272,
      "step": 31570
    },
    {
      "epoch": 16.54874213836478,
      "grad_norm": 0.5866981744766235,
      "learning_rate": 4.401250267041996e-06,
      "loss": 0.5352,
      "num_input_tokens_seen": 20641536,
      "step": 31575
    },
    {
      "epoch": 16.551362683438157,
      "grad_norm": 0.4821137487888336,
      "learning_rate": 4.394773041030137e-06,
      "loss": 0.3634,
      "num_input_tokens_seen": 20644864,
      "step": 31580
    },
    {
      "epoch": 16.55398322851153,
      "grad_norm": 0.3812025785446167,
      "learning_rate": 4.388300125396685e-06,
      "loss": 0.5177,
      "num_input_tokens_seen": 20648224,
      "step": 31585
    },
    {
      "epoch": 16.556603773584907,
      "grad_norm": 0.5001357793807983,
      "learning_rate": 4.381831521495699e-06,
      "loss": 0.4288,
      "num_input_tokens_seen": 20652800,
      "step": 31590
    },
    {
      "epoch": 16.55922431865828,
      "grad_norm": 0.4524155259132385,
      "learning_rate": 4.37536723068033e-06,
      "loss": 0.5388,
      "num_input_tokens_seen": 20655776,
      "step": 31595
    },
    {
      "epoch": 16.561844863731658,
      "grad_norm": 0.3866302967071533,
      "learning_rate": 4.368907254302837e-06,
      "loss": 0.5704,
      "num_input_tokens_seen": 20659040,
      "step": 31600
    },
    {
      "epoch": 16.56446540880503,
      "grad_norm": 0.3706270456314087,
      "learning_rate": 4.3624515937145725e-06,
      "loss": 0.4012,
      "num_input_tokens_seen": 20662560,
      "step": 31605
    },
    {
      "epoch": 16.567085953878408,
      "grad_norm": 0.7102674841880798,
      "learning_rate": 4.3560002502659905e-06,
      "loss": 0.31,
      "num_input_tokens_seen": 20664992,
      "step": 31610
    },
    {
      "epoch": 16.56970649895178,
      "grad_norm": 0.34201905131340027,
      "learning_rate": 4.349553225306644e-06,
      "loss": 0.4522,
      "num_input_tokens_seen": 20667488,
      "step": 31615
    },
    {
      "epoch": 16.572327044025158,
      "grad_norm": 0.9393469095230103,
      "learning_rate": 4.343110520185167e-06,
      "loss": 0.4791,
      "num_input_tokens_seen": 20670304,
      "step": 31620
    },
    {
      "epoch": 16.57494758909853,
      "grad_norm": 0.586219310760498,
      "learning_rate": 4.336672136249298e-06,
      "loss": 0.404,
      "num_input_tokens_seen": 20674080,
      "step": 31625
    },
    {
      "epoch": 16.57756813417191,
      "grad_norm": 0.7089035511016846,
      "learning_rate": 4.33023807484588e-06,
      "loss": 0.3893,
      "num_input_tokens_seen": 20676928,
      "step": 31630
    },
    {
      "epoch": 16.580188679245282,
      "grad_norm": 0.3695134222507477,
      "learning_rate": 4.323808337320842e-06,
      "loss": 0.3658,
      "num_input_tokens_seen": 20680096,
      "step": 31635
    },
    {
      "epoch": 16.58280922431866,
      "grad_norm": 0.5139833092689514,
      "learning_rate": 4.31738292501922e-06,
      "loss": 0.5008,
      "num_input_tokens_seen": 20683360,
      "step": 31640
    },
    {
      "epoch": 16.585429769392032,
      "grad_norm": 0.73460853099823,
      "learning_rate": 4.31096183928513e-06,
      "loss": 0.4517,
      "num_input_tokens_seen": 20685920,
      "step": 31645
    },
    {
      "epoch": 16.58805031446541,
      "grad_norm": 0.7721465826034546,
      "learning_rate": 4.304545081461783e-06,
      "loss": 0.3699,
      "num_input_tokens_seen": 20688576,
      "step": 31650
    },
    {
      "epoch": 16.590670859538783,
      "grad_norm": 0.8507891297340393,
      "learning_rate": 4.298132652891501e-06,
      "loss": 0.3918,
      "num_input_tokens_seen": 20691424,
      "step": 31655
    },
    {
      "epoch": 16.59329140461216,
      "grad_norm": 0.464710533618927,
      "learning_rate": 4.291724554915688e-06,
      "loss": 0.4354,
      "num_input_tokens_seen": 20694432,
      "step": 31660
    },
    {
      "epoch": 16.595911949685533,
      "grad_norm": 0.35295698046684265,
      "learning_rate": 4.285320788874853e-06,
      "loss": 0.3915,
      "num_input_tokens_seen": 20697856,
      "step": 31665
    },
    {
      "epoch": 16.59853249475891,
      "grad_norm": 0.5259097218513489,
      "learning_rate": 4.278921356108581e-06,
      "loss": 0.5406,
      "num_input_tokens_seen": 20701504,
      "step": 31670
    },
    {
      "epoch": 16.601153039832283,
      "grad_norm": 0.3844698667526245,
      "learning_rate": 4.272526257955556e-06,
      "loss": 0.4323,
      "num_input_tokens_seen": 20705056,
      "step": 31675
    },
    {
      "epoch": 16.60377358490566,
      "grad_norm": 0.6083951592445374,
      "learning_rate": 4.266135495753567e-06,
      "loss": 0.4123,
      "num_input_tokens_seen": 20707872,
      "step": 31680
    },
    {
      "epoch": 16.606394129979037,
      "grad_norm": 0.35485929250717163,
      "learning_rate": 4.259749070839491e-06,
      "loss": 0.4378,
      "num_input_tokens_seen": 20710560,
      "step": 31685
    },
    {
      "epoch": 16.60901467505241,
      "grad_norm": 0.5609503984451294,
      "learning_rate": 4.253366984549289e-06,
      "loss": 0.4655,
      "num_input_tokens_seen": 20714016,
      "step": 31690
    },
    {
      "epoch": 16.611635220125788,
      "grad_norm": 1.4186989068984985,
      "learning_rate": 4.24698923821803e-06,
      "loss": 0.3978,
      "num_input_tokens_seen": 20717120,
      "step": 31695
    },
    {
      "epoch": 16.61425576519916,
      "grad_norm": 0.37943196296691895,
      "learning_rate": 4.240615833179861e-06,
      "loss": 0.3662,
      "num_input_tokens_seen": 20720160,
      "step": 31700
    },
    {
      "epoch": 16.616876310272538,
      "grad_norm": 0.375928670167923,
      "learning_rate": 4.234246770768016e-06,
      "loss": 0.4626,
      "num_input_tokens_seen": 20723328,
      "step": 31705
    },
    {
      "epoch": 16.61949685534591,
      "grad_norm": 0.5956690907478333,
      "learning_rate": 4.227882052314841e-06,
      "loss": 0.4004,
      "num_input_tokens_seen": 20725632,
      "step": 31710
    },
    {
      "epoch": 16.622117400419288,
      "grad_norm": 0.7641700506210327,
      "learning_rate": 4.2215216791517575e-06,
      "loss": 0.5817,
      "num_input_tokens_seen": 20729600,
      "step": 31715
    },
    {
      "epoch": 16.62473794549266,
      "grad_norm": 0.43825653195381165,
      "learning_rate": 4.215165652609293e-06,
      "loss": 0.4854,
      "num_input_tokens_seen": 20732512,
      "step": 31720
    },
    {
      "epoch": 16.62735849056604,
      "grad_norm": 0.85433429479599,
      "learning_rate": 4.208813974017048e-06,
      "loss": 0.4723,
      "num_input_tokens_seen": 20736544,
      "step": 31725
    },
    {
      "epoch": 16.629979035639412,
      "grad_norm": 1.3033899068832397,
      "learning_rate": 4.202466644703712e-06,
      "loss": 0.5226,
      "num_input_tokens_seen": 20739968,
      "step": 31730
    },
    {
      "epoch": 16.63259958071279,
      "grad_norm": 0.8026071190834045,
      "learning_rate": 4.196123665997087e-06,
      "loss": 0.4455,
      "num_input_tokens_seen": 20743296,
      "step": 31735
    },
    {
      "epoch": 16.635220125786162,
      "grad_norm": 0.41552767157554626,
      "learning_rate": 4.189785039224045e-06,
      "loss": 0.4963,
      "num_input_tokens_seen": 20745696,
      "step": 31740
    },
    {
      "epoch": 16.63784067085954,
      "grad_norm": 0.3969545364379883,
      "learning_rate": 4.183450765710567e-06,
      "loss": 0.5767,
      "num_input_tokens_seen": 20750272,
      "step": 31745
    },
    {
      "epoch": 16.640461215932913,
      "grad_norm": 0.6308606863021851,
      "learning_rate": 4.177120846781693e-06,
      "loss": 0.4337,
      "num_input_tokens_seen": 20754112,
      "step": 31750
    },
    {
      "epoch": 16.64308176100629,
      "grad_norm": 0.3376530408859253,
      "learning_rate": 4.170795283761583e-06,
      "loss": 0.3932,
      "num_input_tokens_seen": 20756928,
      "step": 31755
    },
    {
      "epoch": 16.645702306079663,
      "grad_norm": 0.5574888586997986,
      "learning_rate": 4.164474077973462e-06,
      "loss": 0.5037,
      "num_input_tokens_seen": 20759968,
      "step": 31760
    },
    {
      "epoch": 16.64832285115304,
      "grad_norm": 0.6169781684875488,
      "learning_rate": 4.158157230739659e-06,
      "loss": 0.436,
      "num_input_tokens_seen": 20762560,
      "step": 31765
    },
    {
      "epoch": 16.650943396226417,
      "grad_norm": 0.6071274280548096,
      "learning_rate": 4.151844743381589e-06,
      "loss": 0.4085,
      "num_input_tokens_seen": 20765120,
      "step": 31770
    },
    {
      "epoch": 16.65356394129979,
      "grad_norm": 0.6205252408981323,
      "learning_rate": 4.145536617219756e-06,
      "loss": 0.3872,
      "num_input_tokens_seen": 20767840,
      "step": 31775
    },
    {
      "epoch": 16.656184486373167,
      "grad_norm": 0.455044686794281,
      "learning_rate": 4.139232853573735e-06,
      "loss": 0.456,
      "num_input_tokens_seen": 20771200,
      "step": 31780
    },
    {
      "epoch": 16.65880503144654,
      "grad_norm": 0.8352575898170471,
      "learning_rate": 4.132933453762214e-06,
      "loss": 0.4713,
      "num_input_tokens_seen": 20774336,
      "step": 31785
    },
    {
      "epoch": 16.661425576519918,
      "grad_norm": 0.43108931183815,
      "learning_rate": 4.126638419102946e-06,
      "loss": 0.4881,
      "num_input_tokens_seen": 20777248,
      "step": 31790
    },
    {
      "epoch": 16.66404612159329,
      "grad_norm": 0.4351421892642975,
      "learning_rate": 4.120347750912784e-06,
      "loss": 0.418,
      "num_input_tokens_seen": 20780288,
      "step": 31795
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 0.8815162777900696,
      "learning_rate": 4.1140614505076724e-06,
      "loss": 0.5775,
      "num_input_tokens_seen": 20782848,
      "step": 31800
    },
    {
      "epoch": 16.66928721174004,
      "grad_norm": 0.4162663519382477,
      "learning_rate": 4.10777951920262e-06,
      "loss": 0.4663,
      "num_input_tokens_seen": 20787168,
      "step": 31805
    },
    {
      "epoch": 16.671907756813418,
      "grad_norm": 0.3869137465953827,
      "learning_rate": 4.1015019583117455e-06,
      "loss": 0.554,
      "num_input_tokens_seen": 20790176,
      "step": 31810
    },
    {
      "epoch": 16.67452830188679,
      "grad_norm": 0.39330658316612244,
      "learning_rate": 4.095228769148232e-06,
      "loss": 0.3975,
      "num_input_tokens_seen": 20792864,
      "step": 31815
    },
    {
      "epoch": 16.67714884696017,
      "grad_norm": 0.45095059275627136,
      "learning_rate": 4.088959953024368e-06,
      "loss": 0.4108,
      "num_input_tokens_seen": 20797408,
      "step": 31820
    },
    {
      "epoch": 16.679769392033542,
      "grad_norm": 0.5459447503089905,
      "learning_rate": 4.082695511251522e-06,
      "loss": 0.3883,
      "num_input_tokens_seen": 20800128,
      "step": 31825
    },
    {
      "epoch": 16.68238993710692,
      "grad_norm": 0.7014859914779663,
      "learning_rate": 4.076435445140131e-06,
      "loss": 0.4899,
      "num_input_tokens_seen": 20803008,
      "step": 31830
    },
    {
      "epoch": 16.685010482180292,
      "grad_norm": 0.4412609040737152,
      "learning_rate": 4.070179755999737e-06,
      "loss": 0.5439,
      "num_input_tokens_seen": 20806144,
      "step": 31835
    },
    {
      "epoch": 16.68763102725367,
      "grad_norm": 0.728438138961792,
      "learning_rate": 4.063928445138962e-06,
      "loss": 0.3676,
      "num_input_tokens_seen": 20809472,
      "step": 31840
    },
    {
      "epoch": 16.690251572327043,
      "grad_norm": 0.6526561379432678,
      "learning_rate": 4.057681513865499e-06,
      "loss": 0.3712,
      "num_input_tokens_seen": 20812288,
      "step": 31845
    },
    {
      "epoch": 16.69287211740042,
      "grad_norm": 0.593073308467865,
      "learning_rate": 4.0514389634861426e-06,
      "loss": 0.4759,
      "num_input_tokens_seen": 20815296,
      "step": 31850
    },
    {
      "epoch": 16.695492662473793,
      "grad_norm": 0.45927202701568604,
      "learning_rate": 4.045200795306767e-06,
      "loss": 0.4914,
      "num_input_tokens_seen": 20822656,
      "step": 31855
    },
    {
      "epoch": 16.69811320754717,
      "grad_norm": 0.31058913469314575,
      "learning_rate": 4.038967010632314e-06,
      "loss": 0.4792,
      "num_input_tokens_seen": 20826016,
      "step": 31860
    },
    {
      "epoch": 16.700733752620543,
      "grad_norm": 0.4337092638015747,
      "learning_rate": 4.032737610766826e-06,
      "loss": 0.3789,
      "num_input_tokens_seen": 20829824,
      "step": 31865
    },
    {
      "epoch": 16.70335429769392,
      "grad_norm": 0.4213986396789551,
      "learning_rate": 4.026512597013429e-06,
      "loss": 0.402,
      "num_input_tokens_seen": 20832512,
      "step": 31870
    },
    {
      "epoch": 16.705974842767297,
      "grad_norm": 0.45716366171836853,
      "learning_rate": 4.020291970674312e-06,
      "loss": 0.4138,
      "num_input_tokens_seen": 20836000,
      "step": 31875
    },
    {
      "epoch": 16.70859538784067,
      "grad_norm": 0.5347204804420471,
      "learning_rate": 4.0140757330507705e-06,
      "loss": 0.2831,
      "num_input_tokens_seen": 20838976,
      "step": 31880
    },
    {
      "epoch": 16.711215932914047,
      "grad_norm": 0.6133154630661011,
      "learning_rate": 4.007863885443161e-06,
      "loss": 0.5378,
      "num_input_tokens_seen": 20842208,
      "step": 31885
    },
    {
      "epoch": 16.71383647798742,
      "grad_norm": 0.5293045043945312,
      "learning_rate": 4.001656429150935e-06,
      "loss": 0.4598,
      "num_input_tokens_seen": 20846272,
      "step": 31890
    },
    {
      "epoch": 16.716457023060798,
      "grad_norm": 0.6392467617988586,
      "learning_rate": 3.9954533654726235e-06,
      "loss": 0.4878,
      "num_input_tokens_seen": 20848896,
      "step": 31895
    },
    {
      "epoch": 16.71907756813417,
      "grad_norm": 0.49942973256111145,
      "learning_rate": 3.989254695705841e-06,
      "loss": 0.4536,
      "num_input_tokens_seen": 20852288,
      "step": 31900
    },
    {
      "epoch": 16.721698113207548,
      "grad_norm": 0.3880143165588379,
      "learning_rate": 3.983060421147272e-06,
      "loss": 0.4797,
      "num_input_tokens_seen": 20855136,
      "step": 31905
    },
    {
      "epoch": 16.72431865828092,
      "grad_norm": 0.4535748064517975,
      "learning_rate": 3.976870543092684e-06,
      "loss": 0.3732,
      "num_input_tokens_seen": 20858432,
      "step": 31910
    },
    {
      "epoch": 16.7269392033543,
      "grad_norm": 0.35663101077079773,
      "learning_rate": 3.970685062836932e-06,
      "loss": 0.3425,
      "num_input_tokens_seen": 20861280,
      "step": 31915
    },
    {
      "epoch": 16.729559748427672,
      "grad_norm": 0.7245160341262817,
      "learning_rate": 3.96450398167395e-06,
      "loss": 0.3906,
      "num_input_tokens_seen": 20863968,
      "step": 31920
    },
    {
      "epoch": 16.73218029350105,
      "grad_norm": 0.784916877746582,
      "learning_rate": 3.958327300896758e-06,
      "loss": 0.5333,
      "num_input_tokens_seen": 20866112,
      "step": 31925
    },
    {
      "epoch": 16.734800838574422,
      "grad_norm": 0.3849329352378845,
      "learning_rate": 3.952155021797429e-06,
      "loss": 0.373,
      "num_input_tokens_seen": 20869856,
      "step": 31930
    },
    {
      "epoch": 16.7374213836478,
      "grad_norm": 0.4248344302177429,
      "learning_rate": 3.945987145667151e-06,
      "loss": 0.411,
      "num_input_tokens_seen": 20872448,
      "step": 31935
    },
    {
      "epoch": 16.740041928721173,
      "grad_norm": 0.2976190447807312,
      "learning_rate": 3.939823673796162e-06,
      "loss": 0.3351,
      "num_input_tokens_seen": 20875008,
      "step": 31940
    },
    {
      "epoch": 16.74266247379455,
      "grad_norm": 0.43316715955734253,
      "learning_rate": 3.933664607473789e-06,
      "loss": 0.5554,
      "num_input_tokens_seen": 20878144,
      "step": 31945
    },
    {
      "epoch": 16.745283018867923,
      "grad_norm": 0.31623852252960205,
      "learning_rate": 3.927509947988445e-06,
      "loss": 0.431,
      "num_input_tokens_seen": 20881216,
      "step": 31950
    },
    {
      "epoch": 16.7479035639413,
      "grad_norm": 0.5698292255401611,
      "learning_rate": 3.9213596966276164e-06,
      "loss": 0.3759,
      "num_input_tokens_seen": 20885120,
      "step": 31955
    },
    {
      "epoch": 16.750524109014677,
      "grad_norm": 0.34228864312171936,
      "learning_rate": 3.9152138546778625e-06,
      "loss": 0.5436,
      "num_input_tokens_seen": 20887808,
      "step": 31960
    },
    {
      "epoch": 16.75314465408805,
      "grad_norm": 0.47411561012268066,
      "learning_rate": 3.9090724234248175e-06,
      "loss": 0.4185,
      "num_input_tokens_seen": 20890720,
      "step": 31965
    },
    {
      "epoch": 16.755765199161427,
      "grad_norm": 0.40852442383766174,
      "learning_rate": 3.902935404153202e-06,
      "loss": 0.3965,
      "num_input_tokens_seen": 20893440,
      "step": 31970
    },
    {
      "epoch": 16.7583857442348,
      "grad_norm": 0.3900405764579773,
      "learning_rate": 3.896802798146809e-06,
      "loss": 0.3903,
      "num_input_tokens_seen": 20897792,
      "step": 31975
    },
    {
      "epoch": 16.761006289308177,
      "grad_norm": 0.4543166756629944,
      "learning_rate": 3.890674606688524e-06,
      "loss": 0.4733,
      "num_input_tokens_seen": 20899936,
      "step": 31980
    },
    {
      "epoch": 16.76362683438155,
      "grad_norm": 0.5250771045684814,
      "learning_rate": 3.88455083106028e-06,
      "loss": 0.395,
      "num_input_tokens_seen": 20905728,
      "step": 31985
    },
    {
      "epoch": 16.766247379454928,
      "grad_norm": 0.5210963487625122,
      "learning_rate": 3.878431472543098e-06,
      "loss": 0.4708,
      "num_input_tokens_seen": 20909376,
      "step": 31990
    },
    {
      "epoch": 16.7688679245283,
      "grad_norm": 0.5270918607711792,
      "learning_rate": 3.872316532417086e-06,
      "loss": 0.4462,
      "num_input_tokens_seen": 20913088,
      "step": 31995
    },
    {
      "epoch": 16.771488469601678,
      "grad_norm": 0.4187776446342468,
      "learning_rate": 3.866206011961418e-06,
      "loss": 0.5471,
      "num_input_tokens_seen": 20915808,
      "step": 32000
    },
    {
      "epoch": 16.77410901467505,
      "grad_norm": 0.381447970867157,
      "learning_rate": 3.860099912454346e-06,
      "loss": 0.5416,
      "num_input_tokens_seen": 20920032,
      "step": 32005
    },
    {
      "epoch": 16.77672955974843,
      "grad_norm": 0.6532714366912842,
      "learning_rate": 3.853998235173201e-06,
      "loss": 0.4076,
      "num_input_tokens_seen": 20922912,
      "step": 32010
    },
    {
      "epoch": 16.779350104821802,
      "grad_norm": 0.6453784108161926,
      "learning_rate": 3.847900981394379e-06,
      "loss": 0.5065,
      "num_input_tokens_seen": 20926176,
      "step": 32015
    },
    {
      "epoch": 16.78197064989518,
      "grad_norm": 0.5820701718330383,
      "learning_rate": 3.841808152393348e-06,
      "loss": 0.4441,
      "num_input_tokens_seen": 20928896,
      "step": 32020
    },
    {
      "epoch": 16.784591194968552,
      "grad_norm": 0.5074016451835632,
      "learning_rate": 3.835719749444666e-06,
      "loss": 0.3878,
      "num_input_tokens_seen": 20931968,
      "step": 32025
    },
    {
      "epoch": 16.78721174004193,
      "grad_norm": 0.5578876733779907,
      "learning_rate": 3.8296357738219606e-06,
      "loss": 0.4498,
      "num_input_tokens_seen": 20935168,
      "step": 32030
    },
    {
      "epoch": 16.789832285115303,
      "grad_norm": 0.37417811155319214,
      "learning_rate": 3.823556226797931e-06,
      "loss": 0.4594,
      "num_input_tokens_seen": 20937856,
      "step": 32035
    },
    {
      "epoch": 16.79245283018868,
      "grad_norm": 0.38141068816185,
      "learning_rate": 3.817481109644344e-06,
      "loss": 0.4393,
      "num_input_tokens_seen": 20941120,
      "step": 32040
    },
    {
      "epoch": 16.795073375262053,
      "grad_norm": 0.5681408047676086,
      "learning_rate": 3.811410423632042e-06,
      "loss": 0.502,
      "num_input_tokens_seen": 20944416,
      "step": 32045
    },
    {
      "epoch": 16.79769392033543,
      "grad_norm": 0.4163997173309326,
      "learning_rate": 3.8053441700309473e-06,
      "loss": 0.4245,
      "num_input_tokens_seen": 20948384,
      "step": 32050
    },
    {
      "epoch": 16.800314465408803,
      "grad_norm": 0.5140409469604492,
      "learning_rate": 3.7992823501100507e-06,
      "loss": 0.4605,
      "num_input_tokens_seen": 20951328,
      "step": 32055
    },
    {
      "epoch": 16.80293501048218,
      "grad_norm": 0.5666776299476624,
      "learning_rate": 3.7932249651374207e-06,
      "loss": 0.4746,
      "num_input_tokens_seen": 20954304,
      "step": 32060
    },
    {
      "epoch": 16.805555555555557,
      "grad_norm": 0.43699437379837036,
      "learning_rate": 3.787172016380183e-06,
      "loss": 0.4129,
      "num_input_tokens_seen": 20957600,
      "step": 32065
    },
    {
      "epoch": 16.80817610062893,
      "grad_norm": 0.285379022359848,
      "learning_rate": 3.7811235051045558e-06,
      "loss": 0.4086,
      "num_input_tokens_seen": 20961248,
      "step": 32070
    },
    {
      "epoch": 16.810796645702307,
      "grad_norm": 0.25222915410995483,
      "learning_rate": 3.775079432575809e-06,
      "loss": 0.3603,
      "num_input_tokens_seen": 20964864,
      "step": 32075
    },
    {
      "epoch": 16.81341719077568,
      "grad_norm": 0.3448941111564636,
      "learning_rate": 3.7690398000582964e-06,
      "loss": 0.4192,
      "num_input_tokens_seen": 20968160,
      "step": 32080
    },
    {
      "epoch": 16.816037735849058,
      "grad_norm": 0.35335254669189453,
      "learning_rate": 3.763004608815443e-06,
      "loss": 0.4543,
      "num_input_tokens_seen": 20971072,
      "step": 32085
    },
    {
      "epoch": 16.81865828092243,
      "grad_norm": 0.4810197055339813,
      "learning_rate": 3.756973860109747e-06,
      "loss": 0.4602,
      "num_input_tokens_seen": 20974016,
      "step": 32090
    },
    {
      "epoch": 16.821278825995808,
      "grad_norm": 0.47600728273391724,
      "learning_rate": 3.7509475552027663e-06,
      "loss": 0.3623,
      "num_input_tokens_seen": 20977280,
      "step": 32095
    },
    {
      "epoch": 16.82389937106918,
      "grad_norm": 0.5784155130386353,
      "learning_rate": 3.744925695355131e-06,
      "loss": 0.5931,
      "num_input_tokens_seen": 20980704,
      "step": 32100
    },
    {
      "epoch": 16.82651991614256,
      "grad_norm": 0.4304869472980499,
      "learning_rate": 3.7389082818265476e-06,
      "loss": 0.6649,
      "num_input_tokens_seen": 20983552,
      "step": 32105
    },
    {
      "epoch": 16.829140461215932,
      "grad_norm": 0.4856566786766052,
      "learning_rate": 3.732895315875792e-06,
      "loss": 0.3929,
      "num_input_tokens_seen": 20986464,
      "step": 32110
    },
    {
      "epoch": 16.83176100628931,
      "grad_norm": 0.5234375,
      "learning_rate": 3.726886798760715e-06,
      "loss": 0.4856,
      "num_input_tokens_seen": 20990112,
      "step": 32115
    },
    {
      "epoch": 16.834381551362682,
      "grad_norm": 0.6311118006706238,
      "learning_rate": 3.720882731738215e-06,
      "loss": 0.4284,
      "num_input_tokens_seen": 20993184,
      "step": 32120
    },
    {
      "epoch": 16.83700209643606,
      "grad_norm": 0.7165728211402893,
      "learning_rate": 3.714883116064291e-06,
      "loss": 0.5249,
      "num_input_tokens_seen": 20995936,
      "step": 32125
    },
    {
      "epoch": 16.839622641509433,
      "grad_norm": 0.7725009322166443,
      "learning_rate": 3.708887952993978e-06,
      "loss": 0.481,
      "num_input_tokens_seen": 20999008,
      "step": 32130
    },
    {
      "epoch": 16.84224318658281,
      "grad_norm": 0.35149985551834106,
      "learning_rate": 3.7028972437814057e-06,
      "loss": 0.4268,
      "num_input_tokens_seen": 21002304,
      "step": 32135
    },
    {
      "epoch": 16.844863731656183,
      "grad_norm": 0.3878001868724823,
      "learning_rate": 3.696910989679764e-06,
      "loss": 0.3834,
      "num_input_tokens_seen": 21005344,
      "step": 32140
    },
    {
      "epoch": 16.84748427672956,
      "grad_norm": 0.3276270925998688,
      "learning_rate": 3.6909291919413003e-06,
      "loss": 0.3789,
      "num_input_tokens_seen": 21009312,
      "step": 32145
    },
    {
      "epoch": 16.850104821802937,
      "grad_norm": 0.9050225615501404,
      "learning_rate": 3.684951851817342e-06,
      "loss": 0.3355,
      "num_input_tokens_seen": 21012544,
      "step": 32150
    },
    {
      "epoch": 16.85272536687631,
      "grad_norm": 0.41128721833229065,
      "learning_rate": 3.678978970558286e-06,
      "loss": 0.3886,
      "num_input_tokens_seen": 21015200,
      "step": 32155
    },
    {
      "epoch": 16.855345911949687,
      "grad_norm": 1.084212303161621,
      "learning_rate": 3.673010549413583e-06,
      "loss": 0.3568,
      "num_input_tokens_seen": 21017696,
      "step": 32160
    },
    {
      "epoch": 16.85796645702306,
      "grad_norm": 0.4668159782886505,
      "learning_rate": 3.6670465896317592e-06,
      "loss": 0.3959,
      "num_input_tokens_seen": 21020800,
      "step": 32165
    },
    {
      "epoch": 16.860587002096437,
      "grad_norm": 0.6025497913360596,
      "learning_rate": 3.661087092460419e-06,
      "loss": 0.3802,
      "num_input_tokens_seen": 21023552,
      "step": 32170
    },
    {
      "epoch": 16.86320754716981,
      "grad_norm": 0.7286758422851562,
      "learning_rate": 3.655132059146202e-06,
      "loss": 0.4405,
      "num_input_tokens_seen": 21026656,
      "step": 32175
    },
    {
      "epoch": 16.865828092243188,
      "grad_norm": 0.44843652844429016,
      "learning_rate": 3.6491814909348526e-06,
      "loss": 0.5109,
      "num_input_tokens_seen": 21030464,
      "step": 32180
    },
    {
      "epoch": 16.86844863731656,
      "grad_norm": 0.4078625738620758,
      "learning_rate": 3.6432353890711424e-06,
      "loss": 0.5003,
      "num_input_tokens_seen": 21033472,
      "step": 32185
    },
    {
      "epoch": 16.871069182389938,
      "grad_norm": 0.34804868698120117,
      "learning_rate": 3.6372937547989425e-06,
      "loss": 0.3797,
      "num_input_tokens_seen": 21036512,
      "step": 32190
    },
    {
      "epoch": 16.87368972746331,
      "grad_norm": 0.5156066417694092,
      "learning_rate": 3.631356589361173e-06,
      "loss": 0.4301,
      "num_input_tokens_seen": 21039936,
      "step": 32195
    },
    {
      "epoch": 16.87631027253669,
      "grad_norm": 0.45716047286987305,
      "learning_rate": 3.625423893999813e-06,
      "loss": 0.5938,
      "num_input_tokens_seen": 21043360,
      "step": 32200
    },
    {
      "epoch": 16.878930817610062,
      "grad_norm": 0.5636375546455383,
      "learning_rate": 3.6194956699559217e-06,
      "loss": 0.4533,
      "num_input_tokens_seen": 21046880,
      "step": 32205
    },
    {
      "epoch": 16.88155136268344,
      "grad_norm": 1.1139110326766968,
      "learning_rate": 3.6135719184696188e-06,
      "loss": 0.2767,
      "num_input_tokens_seen": 21049664,
      "step": 32210
    },
    {
      "epoch": 16.884171907756812,
      "grad_norm": 0.5718989968299866,
      "learning_rate": 3.6076526407800766e-06,
      "loss": 0.4826,
      "num_input_tokens_seen": 21052832,
      "step": 32215
    },
    {
      "epoch": 16.88679245283019,
      "grad_norm": 0.527485191822052,
      "learning_rate": 3.601737838125552e-06,
      "loss": 0.3957,
      "num_input_tokens_seen": 21057120,
      "step": 32220
    },
    {
      "epoch": 16.889412997903563,
      "grad_norm": 0.5625930428504944,
      "learning_rate": 3.595827511743341e-06,
      "loss": 0.5417,
      "num_input_tokens_seen": 21060512,
      "step": 32225
    },
    {
      "epoch": 16.89203354297694,
      "grad_norm": 0.5396862030029297,
      "learning_rate": 3.5899216628698217e-06,
      "loss": 0.443,
      "num_input_tokens_seen": 21062720,
      "step": 32230
    },
    {
      "epoch": 16.894654088050313,
      "grad_norm": 0.37903931736946106,
      "learning_rate": 3.584020292740431e-06,
      "loss": 0.3821,
      "num_input_tokens_seen": 21065504,
      "step": 32235
    },
    {
      "epoch": 16.89727463312369,
      "grad_norm": 0.7760151624679565,
      "learning_rate": 3.578123402589675e-06,
      "loss": 0.3127,
      "num_input_tokens_seen": 21069472,
      "step": 32240
    },
    {
      "epoch": 16.899895178197063,
      "grad_norm": 0.7400917410850525,
      "learning_rate": 3.5722309936511044e-06,
      "loss": 0.4674,
      "num_input_tokens_seen": 21072192,
      "step": 32245
    },
    {
      "epoch": 16.90251572327044,
      "grad_norm": 0.42453745007514954,
      "learning_rate": 3.5663430671573572e-06,
      "loss": 0.3839,
      "num_input_tokens_seen": 21075264,
      "step": 32250
    },
    {
      "epoch": 16.905136268343817,
      "grad_norm": 0.4861743152141571,
      "learning_rate": 3.5604596243401038e-06,
      "loss": 0.3966,
      "num_input_tokens_seen": 21079008,
      "step": 32255
    },
    {
      "epoch": 16.90775681341719,
      "grad_norm": 0.4914734959602356,
      "learning_rate": 3.554580666430102e-06,
      "loss": 0.3945,
      "num_input_tokens_seen": 21081888,
      "step": 32260
    },
    {
      "epoch": 16.910377358490567,
      "grad_norm": 0.8529842495918274,
      "learning_rate": 3.5487061946571614e-06,
      "loss": 0.47,
      "num_input_tokens_seen": 21084608,
      "step": 32265
    },
    {
      "epoch": 16.91299790356394,
      "grad_norm": 0.3371669352054596,
      "learning_rate": 3.542836210250161e-06,
      "loss": 0.6022,
      "num_input_tokens_seen": 21087680,
      "step": 32270
    },
    {
      "epoch": 16.915618448637318,
      "grad_norm": 0.3413989841938019,
      "learning_rate": 3.536970714437032e-06,
      "loss": 0.4923,
      "num_input_tokens_seen": 21090944,
      "step": 32275
    },
    {
      "epoch": 16.91823899371069,
      "grad_norm": 0.37599000334739685,
      "learning_rate": 3.5311097084447575e-06,
      "loss": 0.3143,
      "num_input_tokens_seen": 21094272,
      "step": 32280
    },
    {
      "epoch": 16.920859538784068,
      "grad_norm": 0.4860684871673584,
      "learning_rate": 3.5252531934994033e-06,
      "loss": 0.4062,
      "num_input_tokens_seen": 21097952,
      "step": 32285
    },
    {
      "epoch": 16.92348008385744,
      "grad_norm": 0.5667181015014648,
      "learning_rate": 3.5194011708260866e-06,
      "loss": 0.4703,
      "num_input_tokens_seen": 21101408,
      "step": 32290
    },
    {
      "epoch": 16.92610062893082,
      "grad_norm": 0.5089441537857056,
      "learning_rate": 3.513553641648984e-06,
      "loss": 0.4891,
      "num_input_tokens_seen": 21105888,
      "step": 32295
    },
    {
      "epoch": 16.928721174004192,
      "grad_norm": 0.6277446150779724,
      "learning_rate": 3.5077106071913307e-06,
      "loss": 0.3921,
      "num_input_tokens_seen": 21108992,
      "step": 32300
    },
    {
      "epoch": 16.93134171907757,
      "grad_norm": 0.4944462776184082,
      "learning_rate": 3.5018720686754133e-06,
      "loss": 0.2991,
      "num_input_tokens_seen": 21111648,
      "step": 32305
    },
    {
      "epoch": 16.933962264150942,
      "grad_norm": 0.5165442824363708,
      "learning_rate": 3.4960380273225985e-06,
      "loss": 0.4949,
      "num_input_tokens_seen": 21115872,
      "step": 32310
    },
    {
      "epoch": 16.93658280922432,
      "grad_norm": 0.6436017155647278,
      "learning_rate": 3.490208484353297e-06,
      "loss": 0.3423,
      "num_input_tokens_seen": 21118880,
      "step": 32315
    },
    {
      "epoch": 16.939203354297693,
      "grad_norm": 0.3750622570514679,
      "learning_rate": 3.4843834409869824e-06,
      "loss": 0.378,
      "num_input_tokens_seen": 21121792,
      "step": 32320
    },
    {
      "epoch": 16.94182389937107,
      "grad_norm": 0.3525528013706207,
      "learning_rate": 3.4785628984421977e-06,
      "loss": 0.7687,
      "num_input_tokens_seen": 21125024,
      "step": 32325
    },
    {
      "epoch": 16.944444444444443,
      "grad_norm": 0.5855938792228699,
      "learning_rate": 3.4727468579365235e-06,
      "loss": 0.5236,
      "num_input_tokens_seen": 21128000,
      "step": 32330
    },
    {
      "epoch": 16.94706498951782,
      "grad_norm": 0.3314478397369385,
      "learning_rate": 3.466935320686604e-06,
      "loss": 0.5507,
      "num_input_tokens_seen": 21131808,
      "step": 32335
    },
    {
      "epoch": 16.949685534591197,
      "grad_norm": 0.6520842909812927,
      "learning_rate": 3.4611282879081553e-06,
      "loss": 0.538,
      "num_input_tokens_seen": 21135488,
      "step": 32340
    },
    {
      "epoch": 16.95230607966457,
      "grad_norm": 0.3795686364173889,
      "learning_rate": 3.455325760815939e-06,
      "loss": 0.4344,
      "num_input_tokens_seen": 21138656,
      "step": 32345
    },
    {
      "epoch": 16.954926624737947,
      "grad_norm": 0.6979913115501404,
      "learning_rate": 3.4495277406237843e-06,
      "loss": 0.3689,
      "num_input_tokens_seen": 21141184,
      "step": 32350
    },
    {
      "epoch": 16.95754716981132,
      "grad_norm": 0.4560841917991638,
      "learning_rate": 3.443734228544565e-06,
      "loss": 0.4391,
      "num_input_tokens_seen": 21144288,
      "step": 32355
    },
    {
      "epoch": 16.960167714884697,
      "grad_norm": 0.48959633708000183,
      "learning_rate": 3.437945225790215e-06,
      "loss": 0.5054,
      "num_input_tokens_seen": 21146912,
      "step": 32360
    },
    {
      "epoch": 16.96278825995807,
      "grad_norm": 0.6177394986152649,
      "learning_rate": 3.432160733571729e-06,
      "loss": 0.3565,
      "num_input_tokens_seen": 21150304,
      "step": 32365
    },
    {
      "epoch": 16.965408805031448,
      "grad_norm": 0.2548593580722809,
      "learning_rate": 3.4263807530991595e-06,
      "loss": 0.4663,
      "num_input_tokens_seen": 21153760,
      "step": 32370
    },
    {
      "epoch": 16.96802935010482,
      "grad_norm": 0.25778302550315857,
      "learning_rate": 3.4206052855816155e-06,
      "loss": 0.4493,
      "num_input_tokens_seen": 21156960,
      "step": 32375
    },
    {
      "epoch": 16.970649895178198,
      "grad_norm": 0.8680140972137451,
      "learning_rate": 3.4148343322272567e-06,
      "loss": 0.4525,
      "num_input_tokens_seen": 21159744,
      "step": 32380
    },
    {
      "epoch": 16.97327044025157,
      "grad_norm": 0.467589408159256,
      "learning_rate": 3.4090678942432923e-06,
      "loss": 0.3617,
      "num_input_tokens_seen": 21166304,
      "step": 32385
    },
    {
      "epoch": 16.97589098532495,
      "grad_norm": 0.5590140223503113,
      "learning_rate": 3.4033059728360018e-06,
      "loss": 0.4078,
      "num_input_tokens_seen": 21169440,
      "step": 32390
    },
    {
      "epoch": 16.978511530398322,
      "grad_norm": 0.49573400616645813,
      "learning_rate": 3.397548569210715e-06,
      "loss": 0.4189,
      "num_input_tokens_seen": 21173696,
      "step": 32395
    },
    {
      "epoch": 16.9811320754717,
      "grad_norm": 0.524817168712616,
      "learning_rate": 3.3917956845718136e-06,
      "loss": 0.4349,
      "num_input_tokens_seen": 21176672,
      "step": 32400
    },
    {
      "epoch": 16.983752620545072,
      "grad_norm": 0.5965291261672974,
      "learning_rate": 3.38604732012274e-06,
      "loss": 0.4108,
      "num_input_tokens_seen": 21180064,
      "step": 32405
    },
    {
      "epoch": 16.98637316561845,
      "grad_norm": 0.8229785561561584,
      "learning_rate": 3.3803034770659825e-06,
      "loss": 0.6558,
      "num_input_tokens_seen": 21183680,
      "step": 32410
    },
    {
      "epoch": 16.988993710691823,
      "grad_norm": 0.639384388923645,
      "learning_rate": 3.3745641566030826e-06,
      "loss": 0.408,
      "num_input_tokens_seen": 21187072,
      "step": 32415
    },
    {
      "epoch": 16.9916142557652,
      "grad_norm": 0.5188774466514587,
      "learning_rate": 3.3688293599346442e-06,
      "loss": 0.5292,
      "num_input_tokens_seen": 21191776,
      "step": 32420
    },
    {
      "epoch": 16.994234800838573,
      "grad_norm": 0.34043964743614197,
      "learning_rate": 3.3630990882603213e-06,
      "loss": 0.5574,
      "num_input_tokens_seen": 21196288,
      "step": 32425
    },
    {
      "epoch": 16.99685534591195,
      "grad_norm": 0.7239513993263245,
      "learning_rate": 3.35737334277883e-06,
      "loss": 0.4041,
      "num_input_tokens_seen": 21200992,
      "step": 32430
    },
    {
      "epoch": 16.999475890985323,
      "grad_norm": 0.5641764998435974,
      "learning_rate": 3.351652124687918e-06,
      "loss": 0.4117,
      "num_input_tokens_seen": 21204320,
      "step": 32435
    },
    {
      "epoch": 17.0,
      "eval_loss": 0.4750930368900299,
      "eval_runtime": 9.764,
      "eval_samples_per_second": 86.85,
      "eval_steps_per_second": 21.712,
      "num_input_tokens_seen": 21204416,
      "step": 32436
    },
    {
      "epoch": 17.0020964360587,
      "grad_norm": 0.4694591164588928,
      "learning_rate": 3.345935435184408e-06,
      "loss": 0.442,
      "num_input_tokens_seen": 21206752,
      "step": 32440
    },
    {
      "epoch": 17.004716981132077,
      "grad_norm": 0.3859940767288208,
      "learning_rate": 3.3402232754641576e-06,
      "loss": 0.3904,
      "num_input_tokens_seen": 21210656,
      "step": 32445
    },
    {
      "epoch": 17.00733752620545,
      "grad_norm": 0.5084306597709656,
      "learning_rate": 3.334515646722089e-06,
      "loss": 0.5311,
      "num_input_tokens_seen": 21213216,
      "step": 32450
    },
    {
      "epoch": 17.009958071278827,
      "grad_norm": 0.42530253529548645,
      "learning_rate": 3.32881255015218e-06,
      "loss": 0.4,
      "num_input_tokens_seen": 21216416,
      "step": 32455
    },
    {
      "epoch": 17.0125786163522,
      "grad_norm": 0.465420663356781,
      "learning_rate": 3.3231139869474433e-06,
      "loss": 0.3222,
      "num_input_tokens_seen": 21220352,
      "step": 32460
    },
    {
      "epoch": 17.015199161425578,
      "grad_norm": 0.6497277617454529,
      "learning_rate": 3.317419958299961e-06,
      "loss": 0.3667,
      "num_input_tokens_seen": 21223296,
      "step": 32465
    },
    {
      "epoch": 17.01781970649895,
      "grad_norm": 0.7042871117591858,
      "learning_rate": 3.31173046540085e-06,
      "loss": 0.3944,
      "num_input_tokens_seen": 21225824,
      "step": 32470
    },
    {
      "epoch": 17.020440251572328,
      "grad_norm": 0.5375025868415833,
      "learning_rate": 3.306045509440295e-06,
      "loss": 0.3512,
      "num_input_tokens_seen": 21228704,
      "step": 32475
    },
    {
      "epoch": 17.0230607966457,
      "grad_norm": 0.6972377300262451,
      "learning_rate": 3.300365091607521e-06,
      "loss": 0.3738,
      "num_input_tokens_seen": 21231648,
      "step": 32480
    },
    {
      "epoch": 17.02568134171908,
      "grad_norm": 0.6555013060569763,
      "learning_rate": 3.2946892130908127e-06,
      "loss": 0.4585,
      "num_input_tokens_seen": 21234048,
      "step": 32485
    },
    {
      "epoch": 17.028301886792452,
      "grad_norm": 0.904420793056488,
      "learning_rate": 3.289017875077488e-06,
      "loss": 0.5107,
      "num_input_tokens_seen": 21237568,
      "step": 32490
    },
    {
      "epoch": 17.03092243186583,
      "grad_norm": 1.0096622705459595,
      "learning_rate": 3.283351078753938e-06,
      "loss": 0.4291,
      "num_input_tokens_seen": 21240672,
      "step": 32495
    },
    {
      "epoch": 17.033542976939202,
      "grad_norm": 0.3685053884983063,
      "learning_rate": 3.2776888253055777e-06,
      "loss": 0.3934,
      "num_input_tokens_seen": 21243200,
      "step": 32500
    },
    {
      "epoch": 17.03616352201258,
      "grad_norm": 0.3136177957057953,
      "learning_rate": 3.272031115916896e-06,
      "loss": 0.3116,
      "num_input_tokens_seen": 21247936,
      "step": 32505
    },
    {
      "epoch": 17.038784067085953,
      "grad_norm": 0.4753464460372925,
      "learning_rate": 3.2663779517714238e-06,
      "loss": 0.4229,
      "num_input_tokens_seen": 21250560,
      "step": 32510
    },
    {
      "epoch": 17.04140461215933,
      "grad_norm": 0.2668328583240509,
      "learning_rate": 3.260729334051729e-06,
      "loss": 0.4718,
      "num_input_tokens_seen": 21254432,
      "step": 32515
    },
    {
      "epoch": 17.044025157232703,
      "grad_norm": 0.5051912665367126,
      "learning_rate": 3.2550852639394446e-06,
      "loss": 0.4282,
      "num_input_tokens_seen": 21257568,
      "step": 32520
    },
    {
      "epoch": 17.04664570230608,
      "grad_norm": 0.5147350430488586,
      "learning_rate": 3.2494457426152452e-06,
      "loss": 0.4675,
      "num_input_tokens_seen": 21260992,
      "step": 32525
    },
    {
      "epoch": 17.049266247379453,
      "grad_norm": 0.6297807693481445,
      "learning_rate": 3.243810771258851e-06,
      "loss": 0.43,
      "num_input_tokens_seen": 21263328,
      "step": 32530
    },
    {
      "epoch": 17.05188679245283,
      "grad_norm": 0.5922480821609497,
      "learning_rate": 3.2381803510490387e-06,
      "loss": 0.4391,
      "num_input_tokens_seen": 21266752,
      "step": 32535
    },
    {
      "epoch": 17.054507337526207,
      "grad_norm": 0.44172152876853943,
      "learning_rate": 3.2325544831636217e-06,
      "loss": 0.4693,
      "num_input_tokens_seen": 21269728,
      "step": 32540
    },
    {
      "epoch": 17.05712788259958,
      "grad_norm": 0.5743797421455383,
      "learning_rate": 3.22693316877947e-06,
      "loss": 0.4644,
      "num_input_tokens_seen": 21272864,
      "step": 32545
    },
    {
      "epoch": 17.059748427672957,
      "grad_norm": 0.5092381238937378,
      "learning_rate": 3.2213164090725057e-06,
      "loss": 0.3709,
      "num_input_tokens_seen": 21276320,
      "step": 32550
    },
    {
      "epoch": 17.06236897274633,
      "grad_norm": 0.5551528930664062,
      "learning_rate": 3.2157042052176763e-06,
      "loss": 0.379,
      "num_input_tokens_seen": 21279360,
      "step": 32555
    },
    {
      "epoch": 17.064989517819708,
      "grad_norm": 1.0292754173278809,
      "learning_rate": 3.2100965583890003e-06,
      "loss": 0.4913,
      "num_input_tokens_seen": 21282112,
      "step": 32560
    },
    {
      "epoch": 17.06761006289308,
      "grad_norm": 0.5454462766647339,
      "learning_rate": 3.204493469759537e-06,
      "loss": 0.5606,
      "num_input_tokens_seen": 21284832,
      "step": 32565
    },
    {
      "epoch": 17.070230607966458,
      "grad_norm": 0.5495477914810181,
      "learning_rate": 3.1988949405013764e-06,
      "loss": 0.4095,
      "num_input_tokens_seen": 21288256,
      "step": 32570
    },
    {
      "epoch": 17.07285115303983,
      "grad_norm": 0.5534763932228088,
      "learning_rate": 3.193300971785676e-06,
      "loss": 0.3859,
      "num_input_tokens_seen": 21291616,
      "step": 32575
    },
    {
      "epoch": 17.07547169811321,
      "grad_norm": 0.91898113489151,
      "learning_rate": 3.1877115647826335e-06,
      "loss": 0.5266,
      "num_input_tokens_seen": 21294720,
      "step": 32580
    },
    {
      "epoch": 17.078092243186582,
      "grad_norm": 0.35126543045043945,
      "learning_rate": 3.1821267206614797e-06,
      "loss": 0.4445,
      "num_input_tokens_seen": 21298400,
      "step": 32585
    },
    {
      "epoch": 17.08071278825996,
      "grad_norm": 0.6167198419570923,
      "learning_rate": 3.176546440590508e-06,
      "loss": 0.4696,
      "num_input_tokens_seen": 21301376,
      "step": 32590
    },
    {
      "epoch": 17.083333333333332,
      "grad_norm": 0.4673968553543091,
      "learning_rate": 3.170970725737041e-06,
      "loss": 0.3667,
      "num_input_tokens_seen": 21303840,
      "step": 32595
    },
    {
      "epoch": 17.08595387840671,
      "grad_norm": 0.7955800294876099,
      "learning_rate": 3.165399577267461e-06,
      "loss": 0.4001,
      "num_input_tokens_seen": 21307136,
      "step": 32600
    },
    {
      "epoch": 17.088574423480082,
      "grad_norm": 0.4450558125972748,
      "learning_rate": 3.159832996347184e-06,
      "loss": 0.3649,
      "num_input_tokens_seen": 21309824,
      "step": 32605
    },
    {
      "epoch": 17.09119496855346,
      "grad_norm": 0.4049035608768463,
      "learning_rate": 3.1542709841406865e-06,
      "loss": 0.3835,
      "num_input_tokens_seen": 21312448,
      "step": 32610
    },
    {
      "epoch": 17.093815513626833,
      "grad_norm": 0.5167812705039978,
      "learning_rate": 3.148713541811468e-06,
      "loss": 0.3508,
      "num_input_tokens_seen": 21315040,
      "step": 32615
    },
    {
      "epoch": 17.09643605870021,
      "grad_norm": 0.544625997543335,
      "learning_rate": 3.1431606705220803e-06,
      "loss": 0.4934,
      "num_input_tokens_seen": 21317696,
      "step": 32620
    },
    {
      "epoch": 17.099056603773583,
      "grad_norm": 0.682030439376831,
      "learning_rate": 3.1376123714341234e-06,
      "loss": 0.5293,
      "num_input_tokens_seen": 21320320,
      "step": 32625
    },
    {
      "epoch": 17.10167714884696,
      "grad_norm": 0.46703076362609863,
      "learning_rate": 3.1320686457082387e-06,
      "loss": 0.2684,
      "num_input_tokens_seen": 21323136,
      "step": 32630
    },
    {
      "epoch": 17.104297693920337,
      "grad_norm": 0.3259012997150421,
      "learning_rate": 3.126529494504113e-06,
      "loss": 0.4613,
      "num_input_tokens_seen": 21325664,
      "step": 32635
    },
    {
      "epoch": 17.10691823899371,
      "grad_norm": 0.3658214211463928,
      "learning_rate": 3.1209949189804744e-06,
      "loss": 0.4041,
      "num_input_tokens_seen": 21328352,
      "step": 32640
    },
    {
      "epoch": 17.109538784067087,
      "grad_norm": 0.6141784191131592,
      "learning_rate": 3.115464920295089e-06,
      "loss": 0.4096,
      "num_input_tokens_seen": 21332704,
      "step": 32645
    },
    {
      "epoch": 17.11215932914046,
      "grad_norm": 0.3881005346775055,
      "learning_rate": 3.1099394996047666e-06,
      "loss": 0.5002,
      "num_input_tokens_seen": 21336352,
      "step": 32650
    },
    {
      "epoch": 17.114779874213838,
      "grad_norm": 0.3156793415546417,
      "learning_rate": 3.104418658065367e-06,
      "loss": 0.3798,
      "num_input_tokens_seen": 21340288,
      "step": 32655
    },
    {
      "epoch": 17.11740041928721,
      "grad_norm": 1.0279673337936401,
      "learning_rate": 3.098902396831782e-06,
      "loss": 0.2895,
      "num_input_tokens_seen": 21344128,
      "step": 32660
    },
    {
      "epoch": 17.120020964360588,
      "grad_norm": 0.40582725405693054,
      "learning_rate": 3.0933907170579646e-06,
      "loss": 0.4497,
      "num_input_tokens_seen": 21347296,
      "step": 32665
    },
    {
      "epoch": 17.12264150943396,
      "grad_norm": 0.6798487901687622,
      "learning_rate": 3.0878836198968835e-06,
      "loss": 0.3986,
      "num_input_tokens_seen": 21350240,
      "step": 32670
    },
    {
      "epoch": 17.12526205450734,
      "grad_norm": 1.3097245693206787,
      "learning_rate": 3.0823811065005564e-06,
      "loss": 0.4214,
      "num_input_tokens_seen": 21352704,
      "step": 32675
    },
    {
      "epoch": 17.127882599580712,
      "grad_norm": 0.6026287078857422,
      "learning_rate": 3.0768831780200542e-06,
      "loss": 0.3831,
      "num_input_tokens_seen": 21355904,
      "step": 32680
    },
    {
      "epoch": 17.13050314465409,
      "grad_norm": 0.4075711965560913,
      "learning_rate": 3.071389835605476e-06,
      "loss": 0.4409,
      "num_input_tokens_seen": 21359136,
      "step": 32685
    },
    {
      "epoch": 17.133123689727462,
      "grad_norm": 0.40632808208465576,
      "learning_rate": 3.065901080405978e-06,
      "loss": 0.512,
      "num_input_tokens_seen": 21362496,
      "step": 32690
    },
    {
      "epoch": 17.13574423480084,
      "grad_norm": 0.39897632598876953,
      "learning_rate": 3.060416913569733e-06,
      "loss": 0.3957,
      "num_input_tokens_seen": 21364928,
      "step": 32695
    },
    {
      "epoch": 17.138364779874212,
      "grad_norm": 0.33285149931907654,
      "learning_rate": 3.054937336243968e-06,
      "loss": 0.4676,
      "num_input_tokens_seen": 21367808,
      "step": 32700
    },
    {
      "epoch": 17.14098532494759,
      "grad_norm": 0.654557466506958,
      "learning_rate": 3.049462349574947e-06,
      "loss": 0.3263,
      "num_input_tokens_seen": 21371040,
      "step": 32705
    },
    {
      "epoch": 17.143605870020963,
      "grad_norm": 0.4953549802303314,
      "learning_rate": 3.043991954707981e-06,
      "loss": 0.4549,
      "num_input_tokens_seen": 21374528,
      "step": 32710
    },
    {
      "epoch": 17.14622641509434,
      "grad_norm": 0.8725064992904663,
      "learning_rate": 3.0385261527874075e-06,
      "loss": 0.3765,
      "num_input_tokens_seen": 21376864,
      "step": 32715
    },
    {
      "epoch": 17.148846960167713,
      "grad_norm": 0.5340457558631897,
      "learning_rate": 3.0330649449566205e-06,
      "loss": 0.4865,
      "num_input_tokens_seen": 21379840,
      "step": 32720
    },
    {
      "epoch": 17.15146750524109,
      "grad_norm": 0.39923301339149475,
      "learning_rate": 3.027608332358034e-06,
      "loss": 0.5553,
      "num_input_tokens_seen": 21383072,
      "step": 32725
    },
    {
      "epoch": 17.154088050314467,
      "grad_norm": 0.44824469089508057,
      "learning_rate": 3.022156316133104e-06,
      "loss": 0.4756,
      "num_input_tokens_seen": 21386080,
      "step": 32730
    },
    {
      "epoch": 17.15670859538784,
      "grad_norm": 0.5544937252998352,
      "learning_rate": 3.016708897422335e-06,
      "loss": 0.4814,
      "num_input_tokens_seen": 21388640,
      "step": 32735
    },
    {
      "epoch": 17.159329140461217,
      "grad_norm": 0.49076294898986816,
      "learning_rate": 3.0112660773652685e-06,
      "loss": 0.4254,
      "num_input_tokens_seen": 21391552,
      "step": 32740
    },
    {
      "epoch": 17.16194968553459,
      "grad_norm": 0.47109493613243103,
      "learning_rate": 3.005827857100482e-06,
      "loss": 0.5121,
      "num_input_tokens_seen": 21395040,
      "step": 32745
    },
    {
      "epoch": 17.164570230607968,
      "grad_norm": 0.5817097425460815,
      "learning_rate": 3.000394237765586e-06,
      "loss": 0.4545,
      "num_input_tokens_seen": 21398688,
      "step": 32750
    },
    {
      "epoch": 17.16719077568134,
      "grad_norm": 0.8243701457977295,
      "learning_rate": 2.9949652204972254e-06,
      "loss": 0.5237,
      "num_input_tokens_seen": 21402144,
      "step": 32755
    },
    {
      "epoch": 17.169811320754718,
      "grad_norm": 0.5638766884803772,
      "learning_rate": 2.989540806431093e-06,
      "loss": 0.6145,
      "num_input_tokens_seen": 21404352,
      "step": 32760
    },
    {
      "epoch": 17.17243186582809,
      "grad_norm": 0.44891583919525146,
      "learning_rate": 2.984120996701917e-06,
      "loss": 0.3225,
      "num_input_tokens_seen": 21407808,
      "step": 32765
    },
    {
      "epoch": 17.17505241090147,
      "grad_norm": 0.5644600987434387,
      "learning_rate": 2.9787057924434618e-06,
      "loss": 0.4966,
      "num_input_tokens_seen": 21410976,
      "step": 32770
    },
    {
      "epoch": 17.177672955974842,
      "grad_norm": 0.3496851921081543,
      "learning_rate": 2.9732951947885166e-06,
      "loss": 0.452,
      "num_input_tokens_seen": 21414976,
      "step": 32775
    },
    {
      "epoch": 17.18029350104822,
      "grad_norm": 0.34288471937179565,
      "learning_rate": 2.967889204868929e-06,
      "loss": 0.3301,
      "num_input_tokens_seen": 21418368,
      "step": 32780
    },
    {
      "epoch": 17.182914046121592,
      "grad_norm": 0.6446732878684998,
      "learning_rate": 2.96248782381556e-06,
      "loss": 0.3051,
      "num_input_tokens_seen": 21421664,
      "step": 32785
    },
    {
      "epoch": 17.18553459119497,
      "grad_norm": 0.26317551732063293,
      "learning_rate": 2.957091052758321e-06,
      "loss": 0.4526,
      "num_input_tokens_seen": 21428544,
      "step": 32790
    },
    {
      "epoch": 17.188155136268342,
      "grad_norm": 0.29459279775619507,
      "learning_rate": 2.9516988928261517e-06,
      "loss": 0.3886,
      "num_input_tokens_seen": 21431520,
      "step": 32795
    },
    {
      "epoch": 17.19077568134172,
      "grad_norm": 0.47282183170318604,
      "learning_rate": 2.9463113451470403e-06,
      "loss": 0.3912,
      "num_input_tokens_seen": 21435104,
      "step": 32800
    },
    {
      "epoch": 17.193396226415093,
      "grad_norm": 0.6400941610336304,
      "learning_rate": 2.94092841084799e-06,
      "loss": 0.4774,
      "num_input_tokens_seen": 21437888,
      "step": 32805
    },
    {
      "epoch": 17.19601677148847,
      "grad_norm": 0.36429253220558167,
      "learning_rate": 2.9355500910550586e-06,
      "loss": 0.4041,
      "num_input_tokens_seen": 21441280,
      "step": 32810
    },
    {
      "epoch": 17.198637316561843,
      "grad_norm": 0.40123099088668823,
      "learning_rate": 2.9301763868933157e-06,
      "loss": 0.3088,
      "num_input_tokens_seen": 21445472,
      "step": 32815
    },
    {
      "epoch": 17.20125786163522,
      "grad_norm": 0.6355294585227966,
      "learning_rate": 2.9248072994868896e-06,
      "loss": 0.4697,
      "num_input_tokens_seen": 21447872,
      "step": 32820
    },
    {
      "epoch": 17.203878406708597,
      "grad_norm": 0.41973865032196045,
      "learning_rate": 2.919442829958932e-06,
      "loss": 0.441,
      "num_input_tokens_seen": 21451456,
      "step": 32825
    },
    {
      "epoch": 17.20649895178197,
      "grad_norm": 0.41680285334587097,
      "learning_rate": 2.9140829794316234e-06,
      "loss": 0.3361,
      "num_input_tokens_seen": 21454560,
      "step": 32830
    },
    {
      "epoch": 17.209119496855347,
      "grad_norm": 0.3704107701778412,
      "learning_rate": 2.908727749026191e-06,
      "loss": 0.4216,
      "num_input_tokens_seen": 21457248,
      "step": 32835
    },
    {
      "epoch": 17.21174004192872,
      "grad_norm": 0.6069259643554688,
      "learning_rate": 2.9033771398628756e-06,
      "loss": 0.3896,
      "num_input_tokens_seen": 21461056,
      "step": 32840
    },
    {
      "epoch": 17.214360587002098,
      "grad_norm": 0.31783998012542725,
      "learning_rate": 2.898031153060973e-06,
      "loss": 0.5396,
      "num_input_tokens_seen": 21466176,
      "step": 32845
    },
    {
      "epoch": 17.21698113207547,
      "grad_norm": 0.8305997848510742,
      "learning_rate": 2.8926897897388033e-06,
      "loss": 0.3616,
      "num_input_tokens_seen": 21469824,
      "step": 32850
    },
    {
      "epoch": 17.219601677148848,
      "grad_norm": 0.549030601978302,
      "learning_rate": 2.8873530510137136e-06,
      "loss": 0.4373,
      "num_input_tokens_seen": 21473472,
      "step": 32855
    },
    {
      "epoch": 17.22222222222222,
      "grad_norm": 0.3979852497577667,
      "learning_rate": 2.8820209380020892e-06,
      "loss": 0.3662,
      "num_input_tokens_seen": 21476640,
      "step": 32860
    },
    {
      "epoch": 17.2248427672956,
      "grad_norm": 0.5133140087127686,
      "learning_rate": 2.8766934518193566e-06,
      "loss": 0.4379,
      "num_input_tokens_seen": 21479712,
      "step": 32865
    },
    {
      "epoch": 17.22746331236897,
      "grad_norm": 0.47911280393600464,
      "learning_rate": 2.87137059357995e-06,
      "loss": 0.4813,
      "num_input_tokens_seen": 21483104,
      "step": 32870
    },
    {
      "epoch": 17.23008385744235,
      "grad_norm": 0.5723851323127747,
      "learning_rate": 2.8660523643973618e-06,
      "loss": 0.4018,
      "num_input_tokens_seen": 21486080,
      "step": 32875
    },
    {
      "epoch": 17.232704402515722,
      "grad_norm": 0.4671342074871063,
      "learning_rate": 2.860738765384105e-06,
      "loss": 0.4025,
      "num_input_tokens_seen": 21489792,
      "step": 32880
    },
    {
      "epoch": 17.2353249475891,
      "grad_norm": 0.48143601417541504,
      "learning_rate": 2.8554297976517154e-06,
      "loss": 0.4139,
      "num_input_tokens_seen": 21492704,
      "step": 32885
    },
    {
      "epoch": 17.237945492662472,
      "grad_norm": 0.3383804261684418,
      "learning_rate": 2.8501254623107707e-06,
      "loss": 0.4649,
      "num_input_tokens_seen": 21496992,
      "step": 32890
    },
    {
      "epoch": 17.24056603773585,
      "grad_norm": 0.5441824197769165,
      "learning_rate": 2.84482576047089e-06,
      "loss": 0.4242,
      "num_input_tokens_seen": 21500320,
      "step": 32895
    },
    {
      "epoch": 17.243186582809223,
      "grad_norm": 0.695220410823822,
      "learning_rate": 2.839530693240694e-06,
      "loss": 0.492,
      "num_input_tokens_seen": 21503584,
      "step": 32900
    },
    {
      "epoch": 17.2458071278826,
      "grad_norm": 0.7305642366409302,
      "learning_rate": 2.8342402617278606e-06,
      "loss": 0.5066,
      "num_input_tokens_seen": 21506240,
      "step": 32905
    },
    {
      "epoch": 17.248427672955973,
      "grad_norm": 0.6250032186508179,
      "learning_rate": 2.82895446703908e-06,
      "loss": 0.3766,
      "num_input_tokens_seen": 21510208,
      "step": 32910
    },
    {
      "epoch": 17.25104821802935,
      "grad_norm": 0.3767373263835907,
      "learning_rate": 2.823673310280084e-06,
      "loss": 0.3438,
      "num_input_tokens_seen": 21513344,
      "step": 32915
    },
    {
      "epoch": 17.253668763102727,
      "grad_norm": 0.5705817341804504,
      "learning_rate": 2.8183967925556305e-06,
      "loss": 0.4515,
      "num_input_tokens_seen": 21515968,
      "step": 32920
    },
    {
      "epoch": 17.2562893081761,
      "grad_norm": 0.39600783586502075,
      "learning_rate": 2.8131249149695115e-06,
      "loss": 0.4393,
      "num_input_tokens_seen": 21519200,
      "step": 32925
    },
    {
      "epoch": 17.258909853249477,
      "grad_norm": 0.5074757933616638,
      "learning_rate": 2.8078576786245397e-06,
      "loss": 0.4819,
      "num_input_tokens_seen": 21522368,
      "step": 32930
    },
    {
      "epoch": 17.26153039832285,
      "grad_norm": 0.40889790654182434,
      "learning_rate": 2.8025950846225557e-06,
      "loss": 0.4013,
      "num_input_tokens_seen": 21525280,
      "step": 32935
    },
    {
      "epoch": 17.264150943396228,
      "grad_norm": 0.610937237739563,
      "learning_rate": 2.7973371340644373e-06,
      "loss": 0.3435,
      "num_input_tokens_seen": 21528736,
      "step": 32940
    },
    {
      "epoch": 17.2667714884696,
      "grad_norm": 0.8074837327003479,
      "learning_rate": 2.7920838280500907e-06,
      "loss": 0.5125,
      "num_input_tokens_seen": 21531904,
      "step": 32945
    },
    {
      "epoch": 17.269392033542978,
      "grad_norm": 0.5817888379096985,
      "learning_rate": 2.7868351676784504e-06,
      "loss": 0.4161,
      "num_input_tokens_seen": 21534592,
      "step": 32950
    },
    {
      "epoch": 17.27201257861635,
      "grad_norm": 0.5057060718536377,
      "learning_rate": 2.7815911540474665e-06,
      "loss": 0.5075,
      "num_input_tokens_seen": 21538880,
      "step": 32955
    },
    {
      "epoch": 17.27463312368973,
      "grad_norm": 0.3416748642921448,
      "learning_rate": 2.7763517882541364e-06,
      "loss": 0.3747,
      "num_input_tokens_seen": 21542688,
      "step": 32960
    },
    {
      "epoch": 17.2772536687631,
      "grad_norm": 0.5514679551124573,
      "learning_rate": 2.771117071394466e-06,
      "loss": 0.4102,
      "num_input_tokens_seen": 21545760,
      "step": 32965
    },
    {
      "epoch": 17.27987421383648,
      "grad_norm": 0.5621619820594788,
      "learning_rate": 2.7658870045635055e-06,
      "loss": 0.4429,
      "num_input_tokens_seen": 21548736,
      "step": 32970
    },
    {
      "epoch": 17.282494758909852,
      "grad_norm": 0.615455150604248,
      "learning_rate": 2.7606615888553234e-06,
      "loss": 0.5539,
      "num_input_tokens_seen": 21551936,
      "step": 32975
    },
    {
      "epoch": 17.28511530398323,
      "grad_norm": 0.8080308437347412,
      "learning_rate": 2.755440825363023e-06,
      "loss": 0.5546,
      "num_input_tokens_seen": 21555296,
      "step": 32980
    },
    {
      "epoch": 17.287735849056602,
      "grad_norm": 0.4522808790206909,
      "learning_rate": 2.75022471517872e-06,
      "loss": 0.4372,
      "num_input_tokens_seen": 21558144,
      "step": 32985
    },
    {
      "epoch": 17.29035639412998,
      "grad_norm": 0.677977442741394,
      "learning_rate": 2.745013259393564e-06,
      "loss": 0.5903,
      "num_input_tokens_seen": 21562048,
      "step": 32990
    },
    {
      "epoch": 17.292976939203353,
      "grad_norm": 0.4398210346698761,
      "learning_rate": 2.739806459097735e-06,
      "loss": 0.4376,
      "num_input_tokens_seen": 21565312,
      "step": 32995
    },
    {
      "epoch": 17.29559748427673,
      "grad_norm": 0.5145267844200134,
      "learning_rate": 2.7346043153804403e-06,
      "loss": 0.4465,
      "num_input_tokens_seen": 21568960,
      "step": 33000
    },
    {
      "epoch": 17.298218029350103,
      "grad_norm": 0.3708707094192505,
      "learning_rate": 2.729406829329903e-06,
      "loss": 0.3964,
      "num_input_tokens_seen": 21572064,
      "step": 33005
    },
    {
      "epoch": 17.30083857442348,
      "grad_norm": 0.34999340772628784,
      "learning_rate": 2.72421400203339e-06,
      "loss": 0.4374,
      "num_input_tokens_seen": 21575328,
      "step": 33010
    },
    {
      "epoch": 17.303459119496857,
      "grad_norm": 0.30089765787124634,
      "learning_rate": 2.719025834577163e-06,
      "loss": 0.4319,
      "num_input_tokens_seen": 21578912,
      "step": 33015
    },
    {
      "epoch": 17.30607966457023,
      "grad_norm": 0.520355224609375,
      "learning_rate": 2.713842328046537e-06,
      "loss": 0.423,
      "num_input_tokens_seen": 21582560,
      "step": 33020
    },
    {
      "epoch": 17.308700209643607,
      "grad_norm": 0.6457391977310181,
      "learning_rate": 2.7086634835258416e-06,
      "loss": 0.4466,
      "num_input_tokens_seen": 21586912,
      "step": 33025
    },
    {
      "epoch": 17.31132075471698,
      "grad_norm": 0.7479308247566223,
      "learning_rate": 2.7034893020984333e-06,
      "loss": 0.4964,
      "num_input_tokens_seen": 21589920,
      "step": 33030
    },
    {
      "epoch": 17.313941299790358,
      "grad_norm": 0.5718863010406494,
      "learning_rate": 2.6983197848466952e-06,
      "loss": 0.4092,
      "num_input_tokens_seen": 21592416,
      "step": 33035
    },
    {
      "epoch": 17.31656184486373,
      "grad_norm": 0.583599865436554,
      "learning_rate": 2.6931549328520265e-06,
      "loss": 0.4966,
      "num_input_tokens_seen": 21596448,
      "step": 33040
    },
    {
      "epoch": 17.319182389937108,
      "grad_norm": 0.37143173813819885,
      "learning_rate": 2.6879947471948523e-06,
      "loss": 0.4491,
      "num_input_tokens_seen": 21599968,
      "step": 33045
    },
    {
      "epoch": 17.32180293501048,
      "grad_norm": 0.30367526412010193,
      "learning_rate": 2.6828392289546285e-06,
      "loss": 0.3536,
      "num_input_tokens_seen": 21603168,
      "step": 33050
    },
    {
      "epoch": 17.32442348008386,
      "grad_norm": 0.39472171664237976,
      "learning_rate": 2.677688379209828e-06,
      "loss": 0.5386,
      "num_input_tokens_seen": 21606336,
      "step": 33055
    },
    {
      "epoch": 17.32704402515723,
      "grad_norm": 0.5041537880897522,
      "learning_rate": 2.6725421990379586e-06,
      "loss": 0.4784,
      "num_input_tokens_seen": 21609728,
      "step": 33060
    },
    {
      "epoch": 17.32966457023061,
      "grad_norm": 0.4830700159072876,
      "learning_rate": 2.667400689515537e-06,
      "loss": 0.4886,
      "num_input_tokens_seen": 21612448,
      "step": 33065
    },
    {
      "epoch": 17.332285115303982,
      "grad_norm": 0.3667480945587158,
      "learning_rate": 2.6622638517181e-06,
      "loss": 0.3859,
      "num_input_tokens_seen": 21615424,
      "step": 33070
    },
    {
      "epoch": 17.33490566037736,
      "grad_norm": 0.8878329396247864,
      "learning_rate": 2.657131686720224e-06,
      "loss": 0.3928,
      "num_input_tokens_seen": 21618400,
      "step": 33075
    },
    {
      "epoch": 17.337526205450732,
      "grad_norm": 0.5449135899543762,
      "learning_rate": 2.652004195595498e-06,
      "loss": 0.482,
      "num_input_tokens_seen": 21622688,
      "step": 33080
    },
    {
      "epoch": 17.34014675052411,
      "grad_norm": 0.653723418712616,
      "learning_rate": 2.6468813794165356e-06,
      "loss": 0.5106,
      "num_input_tokens_seen": 21626176,
      "step": 33085
    },
    {
      "epoch": 17.342767295597483,
      "grad_norm": 1.7250220775604248,
      "learning_rate": 2.641763239254977e-06,
      "loss": 0.324,
      "num_input_tokens_seen": 21629440,
      "step": 33090
    },
    {
      "epoch": 17.34538784067086,
      "grad_norm": 0.4306311309337616,
      "learning_rate": 2.6366497761814713e-06,
      "loss": 0.6013,
      "num_input_tokens_seen": 21632000,
      "step": 33095
    },
    {
      "epoch": 17.348008385744233,
      "grad_norm": 0.2952103614807129,
      "learning_rate": 2.631540991265691e-06,
      "loss": 0.4196,
      "num_input_tokens_seen": 21635104,
      "step": 33100
    },
    {
      "epoch": 17.35062893081761,
      "grad_norm": 0.3562178313732147,
      "learning_rate": 2.6264368855763443e-06,
      "loss": 0.5508,
      "num_input_tokens_seen": 21638624,
      "step": 33105
    },
    {
      "epoch": 17.353249475890987,
      "grad_norm": 0.5218148231506348,
      "learning_rate": 2.6213374601811504e-06,
      "loss": 0.3891,
      "num_input_tokens_seen": 21641664,
      "step": 33110
    },
    {
      "epoch": 17.35587002096436,
      "grad_norm": 0.5322993397712708,
      "learning_rate": 2.616242716146855e-06,
      "loss": 0.4893,
      "num_input_tokens_seen": 21644672,
      "step": 33115
    },
    {
      "epoch": 17.358490566037737,
      "grad_norm": 0.7600917220115662,
      "learning_rate": 2.6111526545392145e-06,
      "loss": 0.4213,
      "num_input_tokens_seen": 21647328,
      "step": 33120
    },
    {
      "epoch": 17.36111111111111,
      "grad_norm": 0.4299909770488739,
      "learning_rate": 2.6060672764230106e-06,
      "loss": 0.3725,
      "num_input_tokens_seen": 21650368,
      "step": 33125
    },
    {
      "epoch": 17.363731656184488,
      "grad_norm": 0.5148312449455261,
      "learning_rate": 2.600986582862047e-06,
      "loss": 0.621,
      "num_input_tokens_seen": 21654400,
      "step": 33130
    },
    {
      "epoch": 17.36635220125786,
      "grad_norm": 0.6071879863739014,
      "learning_rate": 2.5959105749191498e-06,
      "loss": 0.3802,
      "num_input_tokens_seen": 21657856,
      "step": 33135
    },
    {
      "epoch": 17.368972746331238,
      "grad_norm": 0.4446136951446533,
      "learning_rate": 2.590839253656166e-06,
      "loss": 0.3123,
      "num_input_tokens_seen": 21661088,
      "step": 33140
    },
    {
      "epoch": 17.37159329140461,
      "grad_norm": 0.4291098713874817,
      "learning_rate": 2.5857726201339473e-06,
      "loss": 0.3344,
      "num_input_tokens_seen": 21666144,
      "step": 33145
    },
    {
      "epoch": 17.37421383647799,
      "grad_norm": 0.4231949746608734,
      "learning_rate": 2.5807106754123873e-06,
      "loss": 0.4066,
      "num_input_tokens_seen": 21669152,
      "step": 33150
    },
    {
      "epoch": 17.37683438155136,
      "grad_norm": 0.4549349844455719,
      "learning_rate": 2.575653420550378e-06,
      "loss": 0.4274,
      "num_input_tokens_seen": 21671840,
      "step": 33155
    },
    {
      "epoch": 17.37945492662474,
      "grad_norm": 0.5558469891548157,
      "learning_rate": 2.57060085660584e-06,
      "loss": 0.3375,
      "num_input_tokens_seen": 21674400,
      "step": 33160
    },
    {
      "epoch": 17.382075471698112,
      "grad_norm": 0.6536827683448792,
      "learning_rate": 2.565552984635716e-06,
      "loss": 0.3765,
      "num_input_tokens_seen": 21677280,
      "step": 33165
    },
    {
      "epoch": 17.38469601677149,
      "grad_norm": 0.776282548904419,
      "learning_rate": 2.5605098056959652e-06,
      "loss": 0.6128,
      "num_input_tokens_seen": 21680352,
      "step": 33170
    },
    {
      "epoch": 17.387316561844862,
      "grad_norm": 0.37514856457710266,
      "learning_rate": 2.555471320841557e-06,
      "loss": 0.4439,
      "num_input_tokens_seen": 21684768,
      "step": 33175
    },
    {
      "epoch": 17.38993710691824,
      "grad_norm": 0.7885450124740601,
      "learning_rate": 2.55043753112649e-06,
      "loss": 0.6144,
      "num_input_tokens_seen": 21687776,
      "step": 33180
    },
    {
      "epoch": 17.392557651991613,
      "grad_norm": 0.4105129539966583,
      "learning_rate": 2.5454084376037717e-06,
      "loss": 0.3425,
      "num_input_tokens_seen": 21691392,
      "step": 33185
    },
    {
      "epoch": 17.39517819706499,
      "grad_norm": 0.6140671968460083,
      "learning_rate": 2.5403840413254304e-06,
      "loss": 0.5999,
      "num_input_tokens_seen": 21694016,
      "step": 33190
    },
    {
      "epoch": 17.397798742138363,
      "grad_norm": 0.47163161635398865,
      "learning_rate": 2.5353643433425196e-06,
      "loss": 0.4685,
      "num_input_tokens_seen": 21697152,
      "step": 33195
    },
    {
      "epoch": 17.40041928721174,
      "grad_norm": 0.26792585849761963,
      "learning_rate": 2.5303493447050965e-06,
      "loss": 0.5282,
      "num_input_tokens_seen": 21701504,
      "step": 33200
    },
    {
      "epoch": 17.403039832285117,
      "grad_norm": 0.5151377320289612,
      "learning_rate": 2.525339046462244e-06,
      "loss": 0.6308,
      "num_input_tokens_seen": 21704640,
      "step": 33205
    },
    {
      "epoch": 17.40566037735849,
      "grad_norm": 0.33853477239608765,
      "learning_rate": 2.5203334496620544e-06,
      "loss": 0.3482,
      "num_input_tokens_seen": 21708160,
      "step": 33210
    },
    {
      "epoch": 17.408280922431867,
      "grad_norm": 0.4937020242214203,
      "learning_rate": 2.515332555351646e-06,
      "loss": 0.4417,
      "num_input_tokens_seen": 21710976,
      "step": 33215
    },
    {
      "epoch": 17.41090146750524,
      "grad_norm": 0.4363376796245575,
      "learning_rate": 2.510336364577154e-06,
      "loss": 0.6596,
      "num_input_tokens_seen": 21714368,
      "step": 33220
    },
    {
      "epoch": 17.413522012578618,
      "grad_norm": 0.575393557548523,
      "learning_rate": 2.5053448783837143e-06,
      "loss": 0.5558,
      "num_input_tokens_seen": 21716896,
      "step": 33225
    },
    {
      "epoch": 17.41614255765199,
      "grad_norm": 0.5066918730735779,
      "learning_rate": 2.5003580978154917e-06,
      "loss": 0.4493,
      "num_input_tokens_seen": 21719488,
      "step": 33230
    },
    {
      "epoch": 17.418763102725368,
      "grad_norm": 0.5553441047668457,
      "learning_rate": 2.4953760239156716e-06,
      "loss": 0.4409,
      "num_input_tokens_seen": 21724000,
      "step": 33235
    },
    {
      "epoch": 17.42138364779874,
      "grad_norm": 0.7475166320800781,
      "learning_rate": 2.4903986577264366e-06,
      "loss": 0.4271,
      "num_input_tokens_seen": 21727424,
      "step": 33240
    },
    {
      "epoch": 17.42400419287212,
      "grad_norm": 0.46835872530937195,
      "learning_rate": 2.4854260002889986e-06,
      "loss": 0.378,
      "num_input_tokens_seen": 21732032,
      "step": 33245
    },
    {
      "epoch": 17.42662473794549,
      "grad_norm": 0.273048996925354,
      "learning_rate": 2.480458052643586e-06,
      "loss": 0.6251,
      "num_input_tokens_seen": 21735488,
      "step": 33250
    },
    {
      "epoch": 17.42924528301887,
      "grad_norm": 0.6347931027412415,
      "learning_rate": 2.4754948158294293e-06,
      "loss": 0.4549,
      "num_input_tokens_seen": 21737984,
      "step": 33255
    },
    {
      "epoch": 17.431865828092242,
      "grad_norm": 0.3511725664138794,
      "learning_rate": 2.4705362908847814e-06,
      "loss": 0.4788,
      "num_input_tokens_seen": 21741088,
      "step": 33260
    },
    {
      "epoch": 17.43448637316562,
      "grad_norm": 0.2850828766822815,
      "learning_rate": 2.4655824788469172e-06,
      "loss": 0.5169,
      "num_input_tokens_seen": 21744928,
      "step": 33265
    },
    {
      "epoch": 17.437106918238992,
      "grad_norm": 1.0487136840820312,
      "learning_rate": 2.4606333807521087e-06,
      "loss": 0.5503,
      "num_input_tokens_seen": 21748288,
      "step": 33270
    },
    {
      "epoch": 17.43972746331237,
      "grad_norm": 0.5103205442428589,
      "learning_rate": 2.455688997635658e-06,
      "loss": 0.4415,
      "num_input_tokens_seen": 21751968,
      "step": 33275
    },
    {
      "epoch": 17.442348008385743,
      "grad_norm": 0.4013684093952179,
      "learning_rate": 2.4507493305318652e-06,
      "loss": 0.4218,
      "num_input_tokens_seen": 21755648,
      "step": 33280
    },
    {
      "epoch": 17.44496855345912,
      "grad_norm": 0.678965151309967,
      "learning_rate": 2.445814380474057e-06,
      "loss": 0.4273,
      "num_input_tokens_seen": 21759040,
      "step": 33285
    },
    {
      "epoch": 17.447589098532493,
      "grad_norm": 0.5436804890632629,
      "learning_rate": 2.4408841484945684e-06,
      "loss": 0.4044,
      "num_input_tokens_seen": 21762272,
      "step": 33290
    },
    {
      "epoch": 17.45020964360587,
      "grad_norm": 0.8569097518920898,
      "learning_rate": 2.435958635624755e-06,
      "loss": 0.5518,
      "num_input_tokens_seen": 21765600,
      "step": 33295
    },
    {
      "epoch": 17.452830188679247,
      "grad_norm": 0.7007933259010315,
      "learning_rate": 2.4310378428949678e-06,
      "loss": 0.4466,
      "num_input_tokens_seen": 21768032,
      "step": 33300
    },
    {
      "epoch": 17.45545073375262,
      "grad_norm": 0.912609338760376,
      "learning_rate": 2.4261217713345804e-06,
      "loss": 0.5061,
      "num_input_tokens_seen": 21770304,
      "step": 33305
    },
    {
      "epoch": 17.458071278825997,
      "grad_norm": 0.45028606057167053,
      "learning_rate": 2.421210421971981e-06,
      "loss": 0.4211,
      "num_input_tokens_seen": 21774016,
      "step": 33310
    },
    {
      "epoch": 17.46069182389937,
      "grad_norm": 0.4116065502166748,
      "learning_rate": 2.416303795834571e-06,
      "loss": 0.5445,
      "num_input_tokens_seen": 21778048,
      "step": 33315
    },
    {
      "epoch": 17.463312368972748,
      "grad_norm": 0.5842188000679016,
      "learning_rate": 2.4114018939487615e-06,
      "loss": 0.2568,
      "num_input_tokens_seen": 21780736,
      "step": 33320
    },
    {
      "epoch": 17.46593291404612,
      "grad_norm": 0.4994579553604126,
      "learning_rate": 2.4065047173399662e-06,
      "loss": 0.4886,
      "num_input_tokens_seen": 21784704,
      "step": 33325
    },
    {
      "epoch": 17.468553459119498,
      "grad_norm": 0.27347415685653687,
      "learning_rate": 2.4016122670326295e-06,
      "loss": 0.334,
      "num_input_tokens_seen": 21788096,
      "step": 33330
    },
    {
      "epoch": 17.47117400419287,
      "grad_norm": 0.40880560874938965,
      "learning_rate": 2.396724544050186e-06,
      "loss": 0.3453,
      "num_input_tokens_seen": 21791808,
      "step": 33335
    },
    {
      "epoch": 17.47379454926625,
      "grad_norm": 0.4139597713947296,
      "learning_rate": 2.3918415494150944e-06,
      "loss": 0.5934,
      "num_input_tokens_seen": 21795072,
      "step": 33340
    },
    {
      "epoch": 17.47641509433962,
      "grad_norm": 0.4643928110599518,
      "learning_rate": 2.3869632841488245e-06,
      "loss": 0.4234,
      "num_input_tokens_seen": 21798016,
      "step": 33345
    },
    {
      "epoch": 17.479035639413,
      "grad_norm": 0.5806607007980347,
      "learning_rate": 2.382089749271857e-06,
      "loss": 0.3677,
      "num_input_tokens_seen": 21801632,
      "step": 33350
    },
    {
      "epoch": 17.481656184486372,
      "grad_norm": 1.002363681793213,
      "learning_rate": 2.3772209458036737e-06,
      "loss": 0.36,
      "num_input_tokens_seen": 21804416,
      "step": 33355
    },
    {
      "epoch": 17.48427672955975,
      "grad_norm": 0.6471282243728638,
      "learning_rate": 2.372356874762771e-06,
      "loss": 0.4941,
      "num_input_tokens_seen": 21808160,
      "step": 33360
    },
    {
      "epoch": 17.486897274633122,
      "grad_norm": 0.862034022808075,
      "learning_rate": 2.36749753716666e-06,
      "loss": 0.4996,
      "num_input_tokens_seen": 21810976,
      "step": 33365
    },
    {
      "epoch": 17.4895178197065,
      "grad_norm": 0.29757705330848694,
      "learning_rate": 2.3626429340318613e-06,
      "loss": 0.4401,
      "num_input_tokens_seen": 21814880,
      "step": 33370
    },
    {
      "epoch": 17.492138364779873,
      "grad_norm": 0.672198474407196,
      "learning_rate": 2.357793066373901e-06,
      "loss": 0.3702,
      "num_input_tokens_seen": 21817792,
      "step": 33375
    },
    {
      "epoch": 17.49475890985325,
      "grad_norm": 0.533906877040863,
      "learning_rate": 2.352947935207317e-06,
      "loss": 0.3387,
      "num_input_tokens_seen": 21821312,
      "step": 33380
    },
    {
      "epoch": 17.497379454926623,
      "grad_norm": 0.32156458497047424,
      "learning_rate": 2.3481075415456495e-06,
      "loss": 0.4074,
      "num_input_tokens_seen": 21825664,
      "step": 33385
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.38668322563171387,
      "learning_rate": 2.3432718864014576e-06,
      "loss": 0.4824,
      "num_input_tokens_seen": 21829088,
      "step": 33390
    },
    {
      "epoch": 17.502620545073377,
      "grad_norm": 0.4327937960624695,
      "learning_rate": 2.338440970786304e-06,
      "loss": 0.4395,
      "num_input_tokens_seen": 21832000,
      "step": 33395
    },
    {
      "epoch": 17.50524109014675,
      "grad_norm": 0.3288714587688446,
      "learning_rate": 2.3336147957107644e-06,
      "loss": 0.3869,
      "num_input_tokens_seen": 21834816,
      "step": 33400
    },
    {
      "epoch": 17.507861635220127,
      "grad_norm": 0.4688219726085663,
      "learning_rate": 2.3287933621844227e-06,
      "loss": 0.4277,
      "num_input_tokens_seen": 21837184,
      "step": 33405
    },
    {
      "epoch": 17.5104821802935,
      "grad_norm": 0.48909449577331543,
      "learning_rate": 2.3239766712158633e-06,
      "loss": 0.4446,
      "num_input_tokens_seen": 21842080,
      "step": 33410
    },
    {
      "epoch": 17.513102725366878,
      "grad_norm": 0.5631869435310364,
      "learning_rate": 2.319164723812675e-06,
      "loss": 0.4367,
      "num_input_tokens_seen": 21844736,
      "step": 33415
    },
    {
      "epoch": 17.51572327044025,
      "grad_norm": 0.3940800130367279,
      "learning_rate": 2.314357520981472e-06,
      "loss": 0.4596,
      "num_input_tokens_seen": 21848736,
      "step": 33420
    },
    {
      "epoch": 17.518343815513628,
      "grad_norm": 0.6825796365737915,
      "learning_rate": 2.309555063727864e-06,
      "loss": 0.3651,
      "num_input_tokens_seen": 21851808,
      "step": 33425
    },
    {
      "epoch": 17.520964360587,
      "grad_norm": 0.3530598282814026,
      "learning_rate": 2.304757353056475e-06,
      "loss": 0.4144,
      "num_input_tokens_seen": 21854752,
      "step": 33430
    },
    {
      "epoch": 17.52358490566038,
      "grad_norm": 0.5527232885360718,
      "learning_rate": 2.2999643899709267e-06,
      "loss": 0.3605,
      "num_input_tokens_seen": 21858016,
      "step": 33435
    },
    {
      "epoch": 17.52620545073375,
      "grad_norm": 0.38727936148643494,
      "learning_rate": 2.2951761754738456e-06,
      "loss": 0.3216,
      "num_input_tokens_seen": 21860928,
      "step": 33440
    },
    {
      "epoch": 17.52882599580713,
      "grad_norm": 0.5683870911598206,
      "learning_rate": 2.29039271056688e-06,
      "loss": 0.3392,
      "num_input_tokens_seen": 21864416,
      "step": 33445
    },
    {
      "epoch": 17.531446540880502,
      "grad_norm": 0.2795754075050354,
      "learning_rate": 2.285613996250674e-06,
      "loss": 0.4392,
      "num_input_tokens_seen": 21867808,
      "step": 33450
    },
    {
      "epoch": 17.53406708595388,
      "grad_norm": 0.2886399030685425,
      "learning_rate": 2.2808400335248865e-06,
      "loss": 0.5065,
      "num_input_tokens_seen": 21871008,
      "step": 33455
    },
    {
      "epoch": 17.536687631027252,
      "grad_norm": 0.8271732926368713,
      "learning_rate": 2.2760708233881654e-06,
      "loss": 0.3892,
      "num_input_tokens_seen": 21873984,
      "step": 33460
    },
    {
      "epoch": 17.53930817610063,
      "grad_norm": 0.43968814611434937,
      "learning_rate": 2.271306366838183e-06,
      "loss": 0.3962,
      "num_input_tokens_seen": 21877216,
      "step": 33465
    },
    {
      "epoch": 17.541928721174003,
      "grad_norm": 0.42651987075805664,
      "learning_rate": 2.266546664871602e-06,
      "loss": 0.2928,
      "num_input_tokens_seen": 21880544,
      "step": 33470
    },
    {
      "epoch": 17.54454926624738,
      "grad_norm": 0.908129870891571,
      "learning_rate": 2.261791718484102e-06,
      "loss": 0.6754,
      "num_input_tokens_seen": 21883296,
      "step": 33475
    },
    {
      "epoch": 17.547169811320753,
      "grad_norm": 0.5496399998664856,
      "learning_rate": 2.257041528670362e-06,
      "loss": 0.4751,
      "num_input_tokens_seen": 21885888,
      "step": 33480
    },
    {
      "epoch": 17.54979035639413,
      "grad_norm": 0.560748815536499,
      "learning_rate": 2.2522960964240753e-06,
      "loss": 0.4402,
      "num_input_tokens_seen": 21888736,
      "step": 33485
    },
    {
      "epoch": 17.552410901467507,
      "grad_norm": 0.5687937140464783,
      "learning_rate": 2.247555422737924e-06,
      "loss": 0.5565,
      "num_input_tokens_seen": 21891264,
      "step": 33490
    },
    {
      "epoch": 17.55503144654088,
      "grad_norm": 0.4986930191516876,
      "learning_rate": 2.2428195086036024e-06,
      "loss": 0.6807,
      "num_input_tokens_seen": 21893760,
      "step": 33495
    },
    {
      "epoch": 17.557651991614257,
      "grad_norm": 0.4222693145275116,
      "learning_rate": 2.238088355011811e-06,
      "loss": 0.4479,
      "num_input_tokens_seen": 21897280,
      "step": 33500
    },
    {
      "epoch": 17.56027253668763,
      "grad_norm": 0.31261545419692993,
      "learning_rate": 2.2333619629522518e-06,
      "loss": 0.4105,
      "num_input_tokens_seen": 21900640,
      "step": 33505
    },
    {
      "epoch": 17.562893081761008,
      "grad_norm": 0.44190680980682373,
      "learning_rate": 2.2286403334136424e-06,
      "loss": 0.418,
      "num_input_tokens_seen": 21903744,
      "step": 33510
    },
    {
      "epoch": 17.56551362683438,
      "grad_norm": 0.7308292388916016,
      "learning_rate": 2.223923467383679e-06,
      "loss": 0.3794,
      "num_input_tokens_seen": 21905984,
      "step": 33515
    },
    {
      "epoch": 17.568134171907758,
      "grad_norm": 0.2775256037712097,
      "learning_rate": 2.2192113658490886e-06,
      "loss": 0.4552,
      "num_input_tokens_seen": 21910272,
      "step": 33520
    },
    {
      "epoch": 17.57075471698113,
      "grad_norm": 0.5180936455726624,
      "learning_rate": 2.2145040297955776e-06,
      "loss": 0.6105,
      "num_input_tokens_seen": 21913216,
      "step": 33525
    },
    {
      "epoch": 17.57337526205451,
      "grad_norm": 0.6816456913948059,
      "learning_rate": 2.209801460207875e-06,
      "loss": 0.3057,
      "num_input_tokens_seen": 21916128,
      "step": 33530
    },
    {
      "epoch": 17.57599580712788,
      "grad_norm": 0.6307173371315002,
      "learning_rate": 2.2051036580697042e-06,
      "loss": 0.3445,
      "num_input_tokens_seen": 21918592,
      "step": 33535
    },
    {
      "epoch": 17.57861635220126,
      "grad_norm": 0.5379856824874878,
      "learning_rate": 2.2004106243637885e-06,
      "loss": 0.4605,
      "num_input_tokens_seen": 21921216,
      "step": 33540
    },
    {
      "epoch": 17.581236897274632,
      "grad_norm": 0.42063188552856445,
      "learning_rate": 2.1957223600718587e-06,
      "loss": 0.4378,
      "num_input_tokens_seen": 21925056,
      "step": 33545
    },
    {
      "epoch": 17.58385744234801,
      "grad_norm": 0.5307743549346924,
      "learning_rate": 2.1910388661746493e-06,
      "loss": 0.4313,
      "num_input_tokens_seen": 21928320,
      "step": 33550
    },
    {
      "epoch": 17.586477987421382,
      "grad_norm": 0.6105784773826599,
      "learning_rate": 2.186360143651886e-06,
      "loss": 0.4654,
      "num_input_tokens_seen": 21931968,
      "step": 33555
    },
    {
      "epoch": 17.58909853249476,
      "grad_norm": 0.4487876892089844,
      "learning_rate": 2.1816861934823103e-06,
      "loss": 0.4014,
      "num_input_tokens_seen": 21935008,
      "step": 33560
    },
    {
      "epoch": 17.591719077568133,
      "grad_norm": 0.7400752305984497,
      "learning_rate": 2.1770170166436595e-06,
      "loss": 0.4433,
      "num_input_tokens_seen": 21937344,
      "step": 33565
    },
    {
      "epoch": 17.59433962264151,
      "grad_norm": 0.5368969440460205,
      "learning_rate": 2.172352614112669e-06,
      "loss": 0.4279,
      "num_input_tokens_seen": 21939936,
      "step": 33570
    },
    {
      "epoch": 17.596960167714883,
      "grad_norm": 0.767258882522583,
      "learning_rate": 2.167692986865083e-06,
      "loss": 0.4181,
      "num_input_tokens_seen": 21942400,
      "step": 33575
    },
    {
      "epoch": 17.59958071278826,
      "grad_norm": 0.9459590911865234,
      "learning_rate": 2.1630381358756362e-06,
      "loss": 0.4053,
      "num_input_tokens_seen": 21945632,
      "step": 33580
    },
    {
      "epoch": 17.602201257861637,
      "grad_norm": 0.6220621466636658,
      "learning_rate": 2.158388062118072e-06,
      "loss": 0.5358,
      "num_input_tokens_seen": 21949312,
      "step": 33585
    },
    {
      "epoch": 17.60482180293501,
      "grad_norm": 0.3467424213886261,
      "learning_rate": 2.1537427665651393e-06,
      "loss": 0.3848,
      "num_input_tokens_seen": 21955872,
      "step": 33590
    },
    {
      "epoch": 17.607442348008387,
      "grad_norm": 0.40374886989593506,
      "learning_rate": 2.1491022501885726e-06,
      "loss": 0.4678,
      "num_input_tokens_seen": 21959712,
      "step": 33595
    },
    {
      "epoch": 17.61006289308176,
      "grad_norm": 0.3062022030353546,
      "learning_rate": 2.1444665139591174e-06,
      "loss": 0.5618,
      "num_input_tokens_seen": 21963136,
      "step": 33600
    },
    {
      "epoch": 17.612683438155138,
      "grad_norm": 0.4837048649787903,
      "learning_rate": 2.139835558846523e-06,
      "loss": 0.5782,
      "num_input_tokens_seen": 21966912,
      "step": 33605
    },
    {
      "epoch": 17.61530398322851,
      "grad_norm": 0.459917277097702,
      "learning_rate": 2.1352093858195225e-06,
      "loss": 0.4785,
      "num_input_tokens_seen": 21969152,
      "step": 33610
    },
    {
      "epoch": 17.617924528301888,
      "grad_norm": 0.9618279337882996,
      "learning_rate": 2.1305879958458676e-06,
      "loss": 0.5028,
      "num_input_tokens_seen": 21971744,
      "step": 33615
    },
    {
      "epoch": 17.62054507337526,
      "grad_norm": 0.807446300983429,
      "learning_rate": 2.1259713898922894e-06,
      "loss": 0.3444,
      "num_input_tokens_seen": 21974592,
      "step": 33620
    },
    {
      "epoch": 17.62316561844864,
      "grad_norm": 0.5254454016685486,
      "learning_rate": 2.1213595689245386e-06,
      "loss": 0.4339,
      "num_input_tokens_seen": 21978528,
      "step": 33625
    },
    {
      "epoch": 17.62578616352201,
      "grad_norm": 0.8682265281677246,
      "learning_rate": 2.1167525339073513e-06,
      "loss": 0.4307,
      "num_input_tokens_seen": 21981792,
      "step": 33630
    },
    {
      "epoch": 17.62840670859539,
      "grad_norm": 0.7577139139175415,
      "learning_rate": 2.112150285804473e-06,
      "loss": 0.3584,
      "num_input_tokens_seen": 21985344,
      "step": 33635
    },
    {
      "epoch": 17.631027253668762,
      "grad_norm": 0.4102667570114136,
      "learning_rate": 2.107552825578632e-06,
      "loss": 0.5015,
      "num_input_tokens_seen": 21988352,
      "step": 33640
    },
    {
      "epoch": 17.63364779874214,
      "grad_norm": 0.3453722596168518,
      "learning_rate": 2.102960154191572e-06,
      "loss": 0.4145,
      "num_input_tokens_seen": 21992608,
      "step": 33645
    },
    {
      "epoch": 17.636268343815512,
      "grad_norm": 0.24839970469474792,
      "learning_rate": 2.098372272604021e-06,
      "loss": 0.3831,
      "num_input_tokens_seen": 21996160,
      "step": 33650
    },
    {
      "epoch": 17.63888888888889,
      "grad_norm": 0.2659134566783905,
      "learning_rate": 2.0937891817757155e-06,
      "loss": 0.4031,
      "num_input_tokens_seen": 21999712,
      "step": 33655
    },
    {
      "epoch": 17.641509433962263,
      "grad_norm": 0.4428059756755829,
      "learning_rate": 2.089210882665385e-06,
      "loss": 0.389,
      "num_input_tokens_seen": 22002432,
      "step": 33660
    },
    {
      "epoch": 17.64412997903564,
      "grad_norm": 0.7382174134254456,
      "learning_rate": 2.0846373762307635e-06,
      "loss": 0.5861,
      "num_input_tokens_seen": 22004992,
      "step": 33665
    },
    {
      "epoch": 17.646750524109013,
      "grad_norm": 0.3510434627532959,
      "learning_rate": 2.080068663428569e-06,
      "loss": 0.4115,
      "num_input_tokens_seen": 22008480,
      "step": 33670
    },
    {
      "epoch": 17.64937106918239,
      "grad_norm": 0.5566017031669617,
      "learning_rate": 2.075504745214521e-06,
      "loss": 0.442,
      "num_input_tokens_seen": 22011968,
      "step": 33675
    },
    {
      "epoch": 17.651991614255767,
      "grad_norm": 0.6377952098846436,
      "learning_rate": 2.0709456225433437e-06,
      "loss": 0.5821,
      "num_input_tokens_seen": 22015424,
      "step": 33680
    },
    {
      "epoch": 17.65461215932914,
      "grad_norm": 0.37517738342285156,
      "learning_rate": 2.0663912963687554e-06,
      "loss": 0.3849,
      "num_input_tokens_seen": 22019776,
      "step": 33685
    },
    {
      "epoch": 17.657232704402517,
      "grad_norm": 0.5787950158119202,
      "learning_rate": 2.0618417676434713e-06,
      "loss": 0.462,
      "num_input_tokens_seen": 22023072,
      "step": 33690
    },
    {
      "epoch": 17.65985324947589,
      "grad_norm": 0.76264888048172,
      "learning_rate": 2.0572970373191975e-06,
      "loss": 0.4197,
      "num_input_tokens_seen": 22026272,
      "step": 33695
    },
    {
      "epoch": 17.662473794549268,
      "grad_norm": 0.4309139549732208,
      "learning_rate": 2.052757106346634e-06,
      "loss": 0.4799,
      "num_input_tokens_seen": 22029216,
      "step": 33700
    },
    {
      "epoch": 17.66509433962264,
      "grad_norm": 0.4219091832637787,
      "learning_rate": 2.048221975675488e-06,
      "loss": 0.4421,
      "num_input_tokens_seen": 22032416,
      "step": 33705
    },
    {
      "epoch": 17.667714884696018,
      "grad_norm": 0.3692972958087921,
      "learning_rate": 2.043691646254456e-06,
      "loss": 0.3296,
      "num_input_tokens_seen": 22035488,
      "step": 33710
    },
    {
      "epoch": 17.67033542976939,
      "grad_norm": 0.5158258080482483,
      "learning_rate": 2.039166119031233e-06,
      "loss": 0.5269,
      "num_input_tokens_seen": 22039232,
      "step": 33715
    },
    {
      "epoch": 17.67295597484277,
      "grad_norm": 0.6551788449287415,
      "learning_rate": 2.034645394952511e-06,
      "loss": 0.466,
      "num_input_tokens_seen": 22041888,
      "step": 33720
    },
    {
      "epoch": 17.67557651991614,
      "grad_norm": 0.36493587493896484,
      "learning_rate": 2.030129474963971e-06,
      "loss": 0.4576,
      "num_input_tokens_seen": 22044416,
      "step": 33725
    },
    {
      "epoch": 17.67819706498952,
      "grad_norm": 0.21090152859687805,
      "learning_rate": 2.025618360010284e-06,
      "loss": 0.3436,
      "num_input_tokens_seen": 22047552,
      "step": 33730
    },
    {
      "epoch": 17.680817610062892,
      "grad_norm": 0.6904630064964294,
      "learning_rate": 2.0211120510351317e-06,
      "loss": 0.4558,
      "num_input_tokens_seen": 22049856,
      "step": 33735
    },
    {
      "epoch": 17.68343815513627,
      "grad_norm": 0.7607049345970154,
      "learning_rate": 2.0166105489811822e-06,
      "loss": 0.444,
      "num_input_tokens_seen": 22052672,
      "step": 33740
    },
    {
      "epoch": 17.686058700209642,
      "grad_norm": 0.46082213521003723,
      "learning_rate": 2.0121138547901016e-06,
      "loss": 0.5429,
      "num_input_tokens_seen": 22055392,
      "step": 33745
    },
    {
      "epoch": 17.68867924528302,
      "grad_norm": 0.5139240622520447,
      "learning_rate": 2.0076219694025427e-06,
      "loss": 0.4409,
      "num_input_tokens_seen": 22058176,
      "step": 33750
    },
    {
      "epoch": 17.691299790356393,
      "grad_norm": 0.7349427342414856,
      "learning_rate": 2.003134893758152e-06,
      "loss": 0.3672,
      "num_input_tokens_seen": 22061152,
      "step": 33755
    },
    {
      "epoch": 17.69392033542977,
      "grad_norm": 0.5033478736877441,
      "learning_rate": 1.9986526287955805e-06,
      "loss": 0.3527,
      "num_input_tokens_seen": 22064288,
      "step": 33760
    },
    {
      "epoch": 17.696540880503143,
      "grad_norm": 0.5395911335945129,
      "learning_rate": 1.9941751754524644e-06,
      "loss": 0.471,
      "num_input_tokens_seen": 22067008,
      "step": 33765
    },
    {
      "epoch": 17.69916142557652,
      "grad_norm": 0.6352372765541077,
      "learning_rate": 1.9897025346654412e-06,
      "loss": 0.3756,
      "num_input_tokens_seen": 22070400,
      "step": 33770
    },
    {
      "epoch": 17.701781970649897,
      "grad_norm": 0.3496614098548889,
      "learning_rate": 1.985234707370134e-06,
      "loss": 0.5108,
      "num_input_tokens_seen": 22073312,
      "step": 33775
    },
    {
      "epoch": 17.70440251572327,
      "grad_norm": 0.6634722352027893,
      "learning_rate": 1.9807716945011544e-06,
      "loss": 0.4532,
      "num_input_tokens_seen": 22076480,
      "step": 33780
    },
    {
      "epoch": 17.707023060796647,
      "grad_norm": 0.46475860476493835,
      "learning_rate": 1.9763134969921215e-06,
      "loss": 0.378,
      "num_input_tokens_seen": 22079968,
      "step": 33785
    },
    {
      "epoch": 17.70964360587002,
      "grad_norm": 0.4359743297100067,
      "learning_rate": 1.971860115775634e-06,
      "loss": 0.4307,
      "num_input_tokens_seen": 22083136,
      "step": 33790
    },
    {
      "epoch": 17.712264150943398,
      "grad_norm": 0.47242817282676697,
      "learning_rate": 1.967411551783294e-06,
      "loss": 0.4721,
      "num_input_tokens_seen": 22086400,
      "step": 33795
    },
    {
      "epoch": 17.71488469601677,
      "grad_norm": 0.40907660126686096,
      "learning_rate": 1.9629678059456895e-06,
      "loss": 0.5252,
      "num_input_tokens_seen": 22089632,
      "step": 33800
    },
    {
      "epoch": 17.717505241090148,
      "grad_norm": 0.4484829306602478,
      "learning_rate": 1.9585288791924004e-06,
      "loss": 0.4419,
      "num_input_tokens_seen": 22092576,
      "step": 33805
    },
    {
      "epoch": 17.72012578616352,
      "grad_norm": 0.4415816366672516,
      "learning_rate": 1.9540947724519947e-06,
      "loss": 0.3712,
      "num_input_tokens_seen": 22096768,
      "step": 33810
    },
    {
      "epoch": 17.7227463312369,
      "grad_norm": 0.9645841717720032,
      "learning_rate": 1.9496654866520414e-06,
      "loss": 0.5344,
      "num_input_tokens_seen": 22099296,
      "step": 33815
    },
    {
      "epoch": 17.72536687631027,
      "grad_norm": 0.4672519862651825,
      "learning_rate": 1.9452410227190937e-06,
      "loss": 0.4238,
      "num_input_tokens_seen": 22101888,
      "step": 33820
    },
    {
      "epoch": 17.72798742138365,
      "grad_norm": 0.59107506275177,
      "learning_rate": 1.9408213815787087e-06,
      "loss": 0.5485,
      "num_input_tokens_seen": 22105568,
      "step": 33825
    },
    {
      "epoch": 17.730607966457022,
      "grad_norm": 0.5745564699172974,
      "learning_rate": 1.936406564155413e-06,
      "loss": 0.4836,
      "num_input_tokens_seen": 22109248,
      "step": 33830
    },
    {
      "epoch": 17.7332285115304,
      "grad_norm": 0.31129515171051025,
      "learning_rate": 1.9319965713727435e-06,
      "loss": 0.393,
      "num_input_tokens_seen": 22112352,
      "step": 33835
    },
    {
      "epoch": 17.735849056603772,
      "grad_norm": 0.3958883583545685,
      "learning_rate": 1.9275914041532144e-06,
      "loss": 0.4549,
      "num_input_tokens_seen": 22115200,
      "step": 33840
    },
    {
      "epoch": 17.73846960167715,
      "grad_norm": 0.6460082530975342,
      "learning_rate": 1.9231910634183394e-06,
      "loss": 0.4935,
      "num_input_tokens_seen": 22118272,
      "step": 33845
    },
    {
      "epoch": 17.741090146750523,
      "grad_norm": 0.4425881803035736,
      "learning_rate": 1.918795550088623e-06,
      "loss": 0.4403,
      "num_input_tokens_seen": 22121248,
      "step": 33850
    },
    {
      "epoch": 17.7437106918239,
      "grad_norm": 0.4868939220905304,
      "learning_rate": 1.9144048650835496e-06,
      "loss": 0.467,
      "num_input_tokens_seen": 22124416,
      "step": 33855
    },
    {
      "epoch": 17.746331236897273,
      "grad_norm": 0.38053077459335327,
      "learning_rate": 1.9100190093216093e-06,
      "loss": 0.366,
      "num_input_tokens_seen": 22127264,
      "step": 33860
    },
    {
      "epoch": 17.74895178197065,
      "grad_norm": 0.404020220041275,
      "learning_rate": 1.9056379837202627e-06,
      "loss": 0.4392,
      "num_input_tokens_seen": 22130816,
      "step": 33865
    },
    {
      "epoch": 17.751572327044027,
      "grad_norm": 0.41621410846710205,
      "learning_rate": 1.9012617891959766e-06,
      "loss": 0.4627,
      "num_input_tokens_seen": 22135200,
      "step": 33870
    },
    {
      "epoch": 17.7541928721174,
      "grad_norm": 0.6298412680625916,
      "learning_rate": 1.8968904266641996e-06,
      "loss": 0.5566,
      "num_input_tokens_seen": 22138304,
      "step": 33875
    },
    {
      "epoch": 17.756813417190777,
      "grad_norm": 0.47585704922676086,
      "learning_rate": 1.892523897039372e-06,
      "loss": 0.303,
      "num_input_tokens_seen": 22141024,
      "step": 33880
    },
    {
      "epoch": 17.75943396226415,
      "grad_norm": 1.1317449808120728,
      "learning_rate": 1.888162201234922e-06,
      "loss": 0.4625,
      "num_input_tokens_seen": 22144800,
      "step": 33885
    },
    {
      "epoch": 17.762054507337528,
      "grad_norm": 0.8913864493370056,
      "learning_rate": 1.8838053401632666e-06,
      "loss": 0.6363,
      "num_input_tokens_seen": 22148256,
      "step": 33890
    },
    {
      "epoch": 17.7646750524109,
      "grad_norm": 0.29202234745025635,
      "learning_rate": 1.8794533147358074e-06,
      "loss": 0.4901,
      "num_input_tokens_seen": 22151840,
      "step": 33895
    },
    {
      "epoch": 17.767295597484278,
      "grad_norm": 0.4270756244659424,
      "learning_rate": 1.875106125862941e-06,
      "loss": 0.2697,
      "num_input_tokens_seen": 22155616,
      "step": 33900
    },
    {
      "epoch": 17.76991614255765,
      "grad_norm": 0.9545901417732239,
      "learning_rate": 1.870763774454057e-06,
      "loss": 0.4423,
      "num_input_tokens_seen": 22158656,
      "step": 33905
    },
    {
      "epoch": 17.77253668763103,
      "grad_norm": 0.5788499116897583,
      "learning_rate": 1.8664262614175115e-06,
      "loss": 0.4598,
      "num_input_tokens_seen": 22161152,
      "step": 33910
    },
    {
      "epoch": 17.7751572327044,
      "grad_norm": 0.35788238048553467,
      "learning_rate": 1.862093587660671e-06,
      "loss": 0.4601,
      "num_input_tokens_seen": 22164544,
      "step": 33915
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 0.7858768105506897,
      "learning_rate": 1.8577657540898851e-06,
      "loss": 0.5612,
      "num_input_tokens_seen": 22167712,
      "step": 33920
    },
    {
      "epoch": 17.780398322851152,
      "grad_norm": 0.3953438103199005,
      "learning_rate": 1.8534427616104798e-06,
      "loss": 0.4936,
      "num_input_tokens_seen": 22171648,
      "step": 33925
    },
    {
      "epoch": 17.78301886792453,
      "grad_norm": 0.5477003455162048,
      "learning_rate": 1.8491246111267846e-06,
      "loss": 0.3789,
      "num_input_tokens_seen": 22178912,
      "step": 33930
    },
    {
      "epoch": 17.785639412997902,
      "grad_norm": 0.8403475880622864,
      "learning_rate": 1.8448113035420934e-06,
      "loss": 0.6284,
      "num_input_tokens_seen": 22181568,
      "step": 33935
    },
    {
      "epoch": 17.78825995807128,
      "grad_norm": 0.30149781703948975,
      "learning_rate": 1.840502839758712e-06,
      "loss": 0.3858,
      "num_input_tokens_seen": 22185408,
      "step": 33940
    },
    {
      "epoch": 17.790880503144653,
      "grad_norm": 0.4376371502876282,
      "learning_rate": 1.8361992206779226e-06,
      "loss": 0.4378,
      "num_input_tokens_seen": 22188224,
      "step": 33945
    },
    {
      "epoch": 17.79350104821803,
      "grad_norm": 0.7397857904434204,
      "learning_rate": 1.8319004471999856e-06,
      "loss": 0.4471,
      "num_input_tokens_seen": 22190688,
      "step": 33950
    },
    {
      "epoch": 17.796121593291403,
      "grad_norm": 1.1055607795715332,
      "learning_rate": 1.8276065202241592e-06,
      "loss": 0.404,
      "num_input_tokens_seen": 22193184,
      "step": 33955
    },
    {
      "epoch": 17.79874213836478,
      "grad_norm": 1.1948115825653076,
      "learning_rate": 1.8233174406486863e-06,
      "loss": 0.3615,
      "num_input_tokens_seen": 22195520,
      "step": 33960
    },
    {
      "epoch": 17.801362683438157,
      "grad_norm": 0.34782674908638,
      "learning_rate": 1.8190332093707907e-06,
      "loss": 0.4583,
      "num_input_tokens_seen": 22198208,
      "step": 33965
    },
    {
      "epoch": 17.80398322851153,
      "grad_norm": 0.34968242049217224,
      "learning_rate": 1.8147538272866833e-06,
      "loss": 0.4444,
      "num_input_tokens_seen": 22201792,
      "step": 33970
    },
    {
      "epoch": 17.806603773584907,
      "grad_norm": 0.4906170964241028,
      "learning_rate": 1.8104792952915672e-06,
      "loss": 0.4178,
      "num_input_tokens_seen": 22205632,
      "step": 33975
    },
    {
      "epoch": 17.80922431865828,
      "grad_norm": 0.4573074281215668,
      "learning_rate": 1.8062096142796192e-06,
      "loss": 0.356,
      "num_input_tokens_seen": 22208192,
      "step": 33980
    },
    {
      "epoch": 17.811844863731658,
      "grad_norm": 0.5243648886680603,
      "learning_rate": 1.8019447851440163e-06,
      "loss": 0.394,
      "num_input_tokens_seen": 22211584,
      "step": 33985
    },
    {
      "epoch": 17.81446540880503,
      "grad_norm": 0.6155127882957458,
      "learning_rate": 1.7976848087769032e-06,
      "loss": 0.562,
      "num_input_tokens_seen": 22215360,
      "step": 33990
    },
    {
      "epoch": 17.817085953878408,
      "grad_norm": 0.7505918145179749,
      "learning_rate": 1.79342968606942e-06,
      "loss": 0.5472,
      "num_input_tokens_seen": 22218848,
      "step": 33995
    },
    {
      "epoch": 17.81970649895178,
      "grad_norm": 0.3384619355201721,
      "learning_rate": 1.7891794179116933e-06,
      "loss": 0.3607,
      "num_input_tokens_seen": 22222880,
      "step": 34000
    },
    {
      "epoch": 17.822327044025158,
      "grad_norm": 0.3878627419471741,
      "learning_rate": 1.7849340051928316e-06,
      "loss": 0.504,
      "num_input_tokens_seen": 22229120,
      "step": 34005
    },
    {
      "epoch": 17.82494758909853,
      "grad_norm": 0.361701637506485,
      "learning_rate": 1.780693448800927e-06,
      "loss": 0.3842,
      "num_input_tokens_seen": 22232352,
      "step": 34010
    },
    {
      "epoch": 17.82756813417191,
      "grad_norm": 0.3632831275463104,
      "learning_rate": 1.7764577496230478e-06,
      "loss": 0.3603,
      "num_input_tokens_seen": 22237856,
      "step": 34015
    },
    {
      "epoch": 17.830188679245282,
      "grad_norm": 0.5754798054695129,
      "learning_rate": 1.7722269085452602e-06,
      "loss": 0.4828,
      "num_input_tokens_seen": 22240992,
      "step": 34020
    },
    {
      "epoch": 17.83280922431866,
      "grad_norm": 0.5423309206962585,
      "learning_rate": 1.7680009264526038e-06,
      "loss": 0.3961,
      "num_input_tokens_seen": 22243776,
      "step": 34025
    },
    {
      "epoch": 17.835429769392032,
      "grad_norm": 0.6619246006011963,
      "learning_rate": 1.7637798042291126e-06,
      "loss": 0.372,
      "num_input_tokens_seen": 22246592,
      "step": 34030
    },
    {
      "epoch": 17.83805031446541,
      "grad_norm": 0.3090479075908661,
      "learning_rate": 1.7595635427577944e-06,
      "loss": 0.3438,
      "num_input_tokens_seen": 22249664,
      "step": 34035
    },
    {
      "epoch": 17.840670859538783,
      "grad_norm": 0.4423450231552124,
      "learning_rate": 1.7553521429206465e-06,
      "loss": 0.407,
      "num_input_tokens_seen": 22253312,
      "step": 34040
    },
    {
      "epoch": 17.84329140461216,
      "grad_norm": 0.4789349436759949,
      "learning_rate": 1.7511456055986335e-06,
      "loss": 0.5214,
      "num_input_tokens_seen": 22256768,
      "step": 34045
    },
    {
      "epoch": 17.845911949685533,
      "grad_norm": 0.6165305972099304,
      "learning_rate": 1.746943931671724e-06,
      "loss": 0.4377,
      "num_input_tokens_seen": 22260224,
      "step": 34050
    },
    {
      "epoch": 17.84853249475891,
      "grad_norm": 0.42219364643096924,
      "learning_rate": 1.742747122018859e-06,
      "loss": 0.3787,
      "num_input_tokens_seen": 22262912,
      "step": 34055
    },
    {
      "epoch": 17.851153039832283,
      "grad_norm": 0.4941461980342865,
      "learning_rate": 1.738555177517967e-06,
      "loss": 0.2892,
      "num_input_tokens_seen": 22266240,
      "step": 34060
    },
    {
      "epoch": 17.85377358490566,
      "grad_norm": 0.3777964413166046,
      "learning_rate": 1.734368099045952e-06,
      "loss": 0.4991,
      "num_input_tokens_seen": 22269472,
      "step": 34065
    },
    {
      "epoch": 17.856394129979037,
      "grad_norm": 0.6844114065170288,
      "learning_rate": 1.730185887478697e-06,
      "loss": 0.3909,
      "num_input_tokens_seen": 22273184,
      "step": 34070
    },
    {
      "epoch": 17.85901467505241,
      "grad_norm": 0.4641265571117401,
      "learning_rate": 1.7260085436910794e-06,
      "loss": 0.3856,
      "num_input_tokens_seen": 22280032,
      "step": 34075
    },
    {
      "epoch": 17.861635220125788,
      "grad_norm": 0.2982912063598633,
      "learning_rate": 1.7218360685569506e-06,
      "loss": 0.5878,
      "num_input_tokens_seen": 22284128,
      "step": 34080
    },
    {
      "epoch": 17.86425576519916,
      "grad_norm": 0.5664168000221252,
      "learning_rate": 1.7176684629491513e-06,
      "loss": 0.4377,
      "num_input_tokens_seen": 22286880,
      "step": 34085
    },
    {
      "epoch": 17.866876310272538,
      "grad_norm": 0.32189464569091797,
      "learning_rate": 1.7135057277394868e-06,
      "loss": 0.3601,
      "num_input_tokens_seen": 22290016,
      "step": 34090
    },
    {
      "epoch": 17.86949685534591,
      "grad_norm": 0.306787371635437,
      "learning_rate": 1.7093478637987548e-06,
      "loss": 0.5098,
      "num_input_tokens_seen": 22294688,
      "step": 34095
    },
    {
      "epoch": 17.872117400419288,
      "grad_norm": 0.7812941074371338,
      "learning_rate": 1.7051948719967375e-06,
      "loss": 0.4137,
      "num_input_tokens_seen": 22297856,
      "step": 34100
    },
    {
      "epoch": 17.87473794549266,
      "grad_norm": 0.8168955445289612,
      "learning_rate": 1.7010467532021901e-06,
      "loss": 0.4834,
      "num_input_tokens_seen": 22301024,
      "step": 34105
    },
    {
      "epoch": 17.87735849056604,
      "grad_norm": 0.3116346001625061,
      "learning_rate": 1.6969035082828544e-06,
      "loss": 0.4942,
      "num_input_tokens_seen": 22303456,
      "step": 34110
    },
    {
      "epoch": 17.879979035639412,
      "grad_norm": 0.4942702651023865,
      "learning_rate": 1.692765138105451e-06,
      "loss": 0.3701,
      "num_input_tokens_seen": 22306528,
      "step": 34115
    },
    {
      "epoch": 17.88259958071279,
      "grad_norm": 0.46325114369392395,
      "learning_rate": 1.6886316435356792e-06,
      "loss": 0.4242,
      "num_input_tokens_seen": 22309824,
      "step": 34120
    },
    {
      "epoch": 17.885220125786162,
      "grad_norm": 1.2665742635726929,
      "learning_rate": 1.6845030254382138e-06,
      "loss": 0.4708,
      "num_input_tokens_seen": 22313024,
      "step": 34125
    },
    {
      "epoch": 17.88784067085954,
      "grad_norm": 0.5924429297447205,
      "learning_rate": 1.6803792846767197e-06,
      "loss": 0.4461,
      "num_input_tokens_seen": 22315584,
      "step": 34130
    },
    {
      "epoch": 17.890461215932913,
      "grad_norm": 0.48748156428337097,
      "learning_rate": 1.6762604221138316e-06,
      "loss": 0.4519,
      "num_input_tokens_seen": 22318144,
      "step": 34135
    },
    {
      "epoch": 17.89308176100629,
      "grad_norm": 0.9326332211494446,
      "learning_rate": 1.6721464386111774e-06,
      "loss": 0.4126,
      "num_input_tokens_seen": 22320800,
      "step": 34140
    },
    {
      "epoch": 17.895702306079663,
      "grad_norm": 0.4950147271156311,
      "learning_rate": 1.6680373350293516e-06,
      "loss": 0.5984,
      "num_input_tokens_seen": 22324064,
      "step": 34145
    },
    {
      "epoch": 17.89832285115304,
      "grad_norm": 0.9561989307403564,
      "learning_rate": 1.6639331122279223e-06,
      "loss": 0.3591,
      "num_input_tokens_seen": 22327104,
      "step": 34150
    },
    {
      "epoch": 17.900943396226417,
      "grad_norm": 0.6298799514770508,
      "learning_rate": 1.6598337710654583e-06,
      "loss": 0.5166,
      "num_input_tokens_seen": 22330528,
      "step": 34155
    },
    {
      "epoch": 17.90356394129979,
      "grad_norm": 0.9077212810516357,
      "learning_rate": 1.6557393123994874e-06,
      "loss": 0.3927,
      "num_input_tokens_seen": 22333472,
      "step": 34160
    },
    {
      "epoch": 17.906184486373167,
      "grad_norm": 0.6701123118400574,
      "learning_rate": 1.651649737086533e-06,
      "loss": 0.4693,
      "num_input_tokens_seen": 22336192,
      "step": 34165
    },
    {
      "epoch": 17.90880503144654,
      "grad_norm": 0.36323240399360657,
      "learning_rate": 1.6475650459820774e-06,
      "loss": 0.4651,
      "num_input_tokens_seen": 22338720,
      "step": 34170
    },
    {
      "epoch": 17.911425576519918,
      "grad_norm": 0.4211600124835968,
      "learning_rate": 1.643485239940598e-06,
      "loss": 0.2986,
      "num_input_tokens_seen": 22341152,
      "step": 34175
    },
    {
      "epoch": 17.91404612159329,
      "grad_norm": 0.4166525900363922,
      "learning_rate": 1.6394103198155375e-06,
      "loss": 0.4527,
      "num_input_tokens_seen": 22344352,
      "step": 34180
    },
    {
      "epoch": 17.916666666666668,
      "grad_norm": 0.4879406988620758,
      "learning_rate": 1.6353402864593276e-06,
      "loss": 0.2752,
      "num_input_tokens_seen": 22346944,
      "step": 34185
    },
    {
      "epoch": 17.91928721174004,
      "grad_norm": 0.3931511342525482,
      "learning_rate": 1.6312751407233706e-06,
      "loss": 0.4538,
      "num_input_tokens_seen": 22351040,
      "step": 34190
    },
    {
      "epoch": 17.921907756813418,
      "grad_norm": 0.39298924803733826,
      "learning_rate": 1.6272148834580559e-06,
      "loss": 0.4761,
      "num_input_tokens_seen": 22353728,
      "step": 34195
    },
    {
      "epoch": 17.92452830188679,
      "grad_norm": 0.6584159135818481,
      "learning_rate": 1.6231595155127317e-06,
      "loss": 0.4852,
      "num_input_tokens_seen": 22356928,
      "step": 34200
    },
    {
      "epoch": 17.92714884696017,
      "grad_norm": 0.6625511646270752,
      "learning_rate": 1.6191090377357414e-06,
      "loss": 0.4001,
      "num_input_tokens_seen": 22359168,
      "step": 34205
    },
    {
      "epoch": 17.929769392033542,
      "grad_norm": 0.5704612135887146,
      "learning_rate": 1.6150634509743939e-06,
      "loss": 0.3915,
      "num_input_tokens_seen": 22362400,
      "step": 34210
    },
    {
      "epoch": 17.93238993710692,
      "grad_norm": 0.5059871077537537,
      "learning_rate": 1.611022756074984e-06,
      "loss": 0.4128,
      "num_input_tokens_seen": 22365408,
      "step": 34215
    },
    {
      "epoch": 17.935010482180292,
      "grad_norm": 0.31610187888145447,
      "learning_rate": 1.606986953882783e-06,
      "loss": 0.3625,
      "num_input_tokens_seen": 22368768,
      "step": 34220
    },
    {
      "epoch": 17.93763102725367,
      "grad_norm": 0.40049564838409424,
      "learning_rate": 1.6029560452420238e-06,
      "loss": 0.4923,
      "num_input_tokens_seen": 22372128,
      "step": 34225
    },
    {
      "epoch": 17.940251572327043,
      "grad_norm": 0.37846869230270386,
      "learning_rate": 1.5989300309959377e-06,
      "loss": 0.4363,
      "num_input_tokens_seen": 22375520,
      "step": 34230
    },
    {
      "epoch": 17.94287211740042,
      "grad_norm": 0.48755475878715515,
      "learning_rate": 1.5949089119867116e-06,
      "loss": 0.6568,
      "num_input_tokens_seen": 22378656,
      "step": 34235
    },
    {
      "epoch": 17.945492662473793,
      "grad_norm": 0.7772616147994995,
      "learning_rate": 1.5908926890555204e-06,
      "loss": 0.4933,
      "num_input_tokens_seen": 22381888,
      "step": 34240
    },
    {
      "epoch": 17.94811320754717,
      "grad_norm": 0.6003173589706421,
      "learning_rate": 1.5868813630425167e-06,
      "loss": 0.471,
      "num_input_tokens_seen": 22384928,
      "step": 34245
    },
    {
      "epoch": 17.950733752620543,
      "grad_norm": 0.384825736284256,
      "learning_rate": 1.5828749347868183e-06,
      "loss": 0.3432,
      "num_input_tokens_seen": 22388096,
      "step": 34250
    },
    {
      "epoch": 17.95335429769392,
      "grad_norm": 0.4154928922653198,
      "learning_rate": 1.5788734051265268e-06,
      "loss": 0.2799,
      "num_input_tokens_seen": 22393632,
      "step": 34255
    },
    {
      "epoch": 17.955974842767297,
      "grad_norm": 0.4066172242164612,
      "learning_rate": 1.5748767748987231e-06,
      "loss": 0.4705,
      "num_input_tokens_seen": 22396768,
      "step": 34260
    },
    {
      "epoch": 17.95859538784067,
      "grad_norm": 0.5288817286491394,
      "learning_rate": 1.5708850449394436e-06,
      "loss": 0.3934,
      "num_input_tokens_seen": 22399776,
      "step": 34265
    },
    {
      "epoch": 17.961215932914047,
      "grad_norm": 0.5046936273574829,
      "learning_rate": 1.5668982160837176e-06,
      "loss": 0.5607,
      "num_input_tokens_seen": 22403104,
      "step": 34270
    },
    {
      "epoch": 17.96383647798742,
      "grad_norm": 0.7402204275131226,
      "learning_rate": 1.562916289165553e-06,
      "loss": 0.368,
      "num_input_tokens_seen": 22406592,
      "step": 34275
    },
    {
      "epoch": 17.966457023060798,
      "grad_norm": 0.28129860758781433,
      "learning_rate": 1.5589392650179114e-06,
      "loss": 0.3961,
      "num_input_tokens_seen": 22411264,
      "step": 34280
    },
    {
      "epoch": 17.96907756813417,
      "grad_norm": 0.4556896686553955,
      "learning_rate": 1.5549671444727437e-06,
      "loss": 0.4167,
      "num_input_tokens_seen": 22414976,
      "step": 34285
    },
    {
      "epoch": 17.971698113207548,
      "grad_norm": 0.39410674571990967,
      "learning_rate": 1.55099992836098e-06,
      "loss": 0.5516,
      "num_input_tokens_seen": 22418016,
      "step": 34290
    },
    {
      "epoch": 17.97431865828092,
      "grad_norm": 0.45462360978126526,
      "learning_rate": 1.5470376175125061e-06,
      "loss": 0.5172,
      "num_input_tokens_seen": 22421408,
      "step": 34295
    },
    {
      "epoch": 17.9769392033543,
      "grad_norm": 0.44220006465911865,
      "learning_rate": 1.5430802127562038e-06,
      "loss": 0.6317,
      "num_input_tokens_seen": 22423744,
      "step": 34300
    },
    {
      "epoch": 17.979559748427672,
      "grad_norm": 0.5586022734642029,
      "learning_rate": 1.5391277149199023e-06,
      "loss": 0.3983,
      "num_input_tokens_seen": 22428672,
      "step": 34305
    },
    {
      "epoch": 17.98218029350105,
      "grad_norm": 0.3423440456390381,
      "learning_rate": 1.5351801248304292e-06,
      "loss": 0.4617,
      "num_input_tokens_seen": 22431648,
      "step": 34310
    },
    {
      "epoch": 17.984800838574422,
      "grad_norm": 0.6047186255455017,
      "learning_rate": 1.5312374433135767e-06,
      "loss": 0.4422,
      "num_input_tokens_seen": 22434496,
      "step": 34315
    },
    {
      "epoch": 17.9874213836478,
      "grad_norm": 0.3273676335811615,
      "learning_rate": 1.5272996711941018e-06,
      "loss": 0.4114,
      "num_input_tokens_seen": 22438432,
      "step": 34320
    },
    {
      "epoch": 17.990041928721173,
      "grad_norm": 0.3779723644256592,
      "learning_rate": 1.5233668092957482e-06,
      "loss": 0.3968,
      "num_input_tokens_seen": 22441600,
      "step": 34325
    },
    {
      "epoch": 17.99266247379455,
      "grad_norm": 0.49080610275268555,
      "learning_rate": 1.5194388584412217e-06,
      "loss": 0.4125,
      "num_input_tokens_seen": 22444160,
      "step": 34330
    },
    {
      "epoch": 17.995283018867923,
      "grad_norm": 0.5774755477905273,
      "learning_rate": 1.5155158194522013e-06,
      "loss": 0.4874,
      "num_input_tokens_seen": 22447328,
      "step": 34335
    },
    {
      "epoch": 17.9979035639413,
      "grad_norm": 0.4396766722202301,
      "learning_rate": 1.5115976931493498e-06,
      "loss": 0.4239,
      "num_input_tokens_seen": 22450112,
      "step": 34340
    },
    {
      "epoch": 18.0,
      "eval_loss": 0.4751274585723877,
      "eval_runtime": 9.7604,
      "eval_samples_per_second": 86.882,
      "eval_steps_per_second": 21.721,
      "num_input_tokens_seen": 22451928,
      "step": 34344
    },
    {
      "epoch": 18.000524109014677,
      "grad_norm": 0.6876239776611328,
      "learning_rate": 1.5076844803522922e-06,
      "loss": 0.3578,
      "num_input_tokens_seen": 22452408,
      "step": 34345
    },
    {
      "epoch": 18.00314465408805,
      "grad_norm": 0.5844520330429077,
      "learning_rate": 1.5037761818796236e-06,
      "loss": 0.4153,
      "num_input_tokens_seen": 22456184,
      "step": 34350
    },
    {
      "epoch": 18.005765199161427,
      "grad_norm": 0.8748747706413269,
      "learning_rate": 1.4998727985489235e-06,
      "loss": 0.4963,
      "num_input_tokens_seen": 22458968,
      "step": 34355
    },
    {
      "epoch": 18.0083857442348,
      "grad_norm": 0.34674784541130066,
      "learning_rate": 1.4959743311767249e-06,
      "loss": 0.3608,
      "num_input_tokens_seen": 22462392,
      "step": 34360
    },
    {
      "epoch": 18.011006289308177,
      "grad_norm": 0.47926777601242065,
      "learning_rate": 1.4920807805785502e-06,
      "loss": 0.4049,
      "num_input_tokens_seen": 22466360,
      "step": 34365
    },
    {
      "epoch": 18.01362683438155,
      "grad_norm": 0.24144987761974335,
      "learning_rate": 1.4881921475688815e-06,
      "loss": 0.4389,
      "num_input_tokens_seen": 22470104,
      "step": 34370
    },
    {
      "epoch": 18.016247379454928,
      "grad_norm": 0.5628008842468262,
      "learning_rate": 1.4843084329611844e-06,
      "loss": 0.4152,
      "num_input_tokens_seen": 22473144,
      "step": 34375
    },
    {
      "epoch": 18.0188679245283,
      "grad_norm": 0.694631040096283,
      "learning_rate": 1.480429637567879e-06,
      "loss": 0.4751,
      "num_input_tokens_seen": 22475640,
      "step": 34380
    },
    {
      "epoch": 18.021488469601678,
      "grad_norm": 0.8221634030342102,
      "learning_rate": 1.4765557622003684e-06,
      "loss": 0.4162,
      "num_input_tokens_seen": 22478712,
      "step": 34385
    },
    {
      "epoch": 18.02410901467505,
      "grad_norm": 0.41017860174179077,
      "learning_rate": 1.4726868076690186e-06,
      "loss": 0.5308,
      "num_input_tokens_seen": 22481624,
      "step": 34390
    },
    {
      "epoch": 18.02672955974843,
      "grad_norm": 0.3766533136367798,
      "learning_rate": 1.4688227747831767e-06,
      "loss": 0.4324,
      "num_input_tokens_seen": 22484664,
      "step": 34395
    },
    {
      "epoch": 18.029350104821802,
      "grad_norm": 0.3983609080314636,
      "learning_rate": 1.4649636643511573e-06,
      "loss": 0.3955,
      "num_input_tokens_seen": 22487128,
      "step": 34400
    },
    {
      "epoch": 18.03197064989518,
      "grad_norm": 0.9483717679977417,
      "learning_rate": 1.4611094771802364e-06,
      "loss": 0.4732,
      "num_input_tokens_seen": 22489528,
      "step": 34405
    },
    {
      "epoch": 18.034591194968552,
      "grad_norm": 0.3718623220920563,
      "learning_rate": 1.457260214076664e-06,
      "loss": 0.5035,
      "num_input_tokens_seen": 22493560,
      "step": 34410
    },
    {
      "epoch": 18.03721174004193,
      "grad_norm": 0.6037285923957825,
      "learning_rate": 1.4534158758456623e-06,
      "loss": 0.4194,
      "num_input_tokens_seen": 22496408,
      "step": 34415
    },
    {
      "epoch": 18.039832285115303,
      "grad_norm": 0.6840581893920898,
      "learning_rate": 1.4495764632914271e-06,
      "loss": 0.374,
      "num_input_tokens_seen": 22500440,
      "step": 34420
    },
    {
      "epoch": 18.04245283018868,
      "grad_norm": 0.676082193851471,
      "learning_rate": 1.4457419772171215e-06,
      "loss": 0.4733,
      "num_input_tokens_seen": 22503640,
      "step": 34425
    },
    {
      "epoch": 18.045073375262053,
      "grad_norm": 0.5142995119094849,
      "learning_rate": 1.4419124184248733e-06,
      "loss": 0.5107,
      "num_input_tokens_seen": 22507768,
      "step": 34430
    },
    {
      "epoch": 18.04769392033543,
      "grad_norm": 0.5430092215538025,
      "learning_rate": 1.4380877877157834e-06,
      "loss": 0.4428,
      "num_input_tokens_seen": 22511256,
      "step": 34435
    },
    {
      "epoch": 18.050314465408807,
      "grad_norm": 0.6094193458557129,
      "learning_rate": 1.4342680858899143e-06,
      "loss": 0.4856,
      "num_input_tokens_seen": 22514872,
      "step": 34440
    },
    {
      "epoch": 18.05293501048218,
      "grad_norm": 0.6491051912307739,
      "learning_rate": 1.4304533137463134e-06,
      "loss": 0.3459,
      "num_input_tokens_seen": 22518008,
      "step": 34445
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 0.659081757068634,
      "learning_rate": 1.4266434720829808e-06,
      "loss": 0.3921,
      "num_input_tokens_seen": 22520664,
      "step": 34450
    },
    {
      "epoch": 18.05817610062893,
      "grad_norm": 0.5950804948806763,
      "learning_rate": 1.4228385616968987e-06,
      "loss": 0.3352,
      "num_input_tokens_seen": 22523096,
      "step": 34455
    },
    {
      "epoch": 18.060796645702307,
      "grad_norm": 0.6990751624107361,
      "learning_rate": 1.419038583384008e-06,
      "loss": 0.4711,
      "num_input_tokens_seen": 22526616,
      "step": 34460
    },
    {
      "epoch": 18.06341719077568,
      "grad_norm": 0.8329511880874634,
      "learning_rate": 1.4152435379392148e-06,
      "loss": 0.4452,
      "num_input_tokens_seen": 22529560,
      "step": 34465
    },
    {
      "epoch": 18.066037735849058,
      "grad_norm": 0.49898675084114075,
      "learning_rate": 1.411453426156406e-06,
      "loss": 0.4059,
      "num_input_tokens_seen": 22533016,
      "step": 34470
    },
    {
      "epoch": 18.06865828092243,
      "grad_norm": 0.9498148560523987,
      "learning_rate": 1.4076682488284282e-06,
      "loss": 0.5011,
      "num_input_tokens_seen": 22535416,
      "step": 34475
    },
    {
      "epoch": 18.071278825995808,
      "grad_norm": 0.4633922874927521,
      "learning_rate": 1.4038880067470978e-06,
      "loss": 0.3262,
      "num_input_tokens_seen": 22538936,
      "step": 34480
    },
    {
      "epoch": 18.07389937106918,
      "grad_norm": 0.6585336327552795,
      "learning_rate": 1.400112700703196e-06,
      "loss": 0.4015,
      "num_input_tokens_seen": 22541944,
      "step": 34485
    },
    {
      "epoch": 18.07651991614256,
      "grad_norm": 0.511701762676239,
      "learning_rate": 1.3963423314864805e-06,
      "loss": 0.512,
      "num_input_tokens_seen": 22544568,
      "step": 34490
    },
    {
      "epoch": 18.079140461215932,
      "grad_norm": 0.43685004115104675,
      "learning_rate": 1.3925768998856587e-06,
      "loss": 0.4457,
      "num_input_tokens_seen": 22547768,
      "step": 34495
    },
    {
      "epoch": 18.08176100628931,
      "grad_norm": 0.5352755188941956,
      "learning_rate": 1.388816406688423e-06,
      "loss": 0.4348,
      "num_input_tokens_seen": 22550776,
      "step": 34500
    },
    {
      "epoch": 18.084381551362682,
      "grad_norm": 0.2896326184272766,
      "learning_rate": 1.3850608526814246e-06,
      "loss": 0.3476,
      "num_input_tokens_seen": 22554392,
      "step": 34505
    },
    {
      "epoch": 18.08700209643606,
      "grad_norm": 0.48258450627326965,
      "learning_rate": 1.3813102386502853e-06,
      "loss": 0.49,
      "num_input_tokens_seen": 22557304,
      "step": 34510
    },
    {
      "epoch": 18.089622641509433,
      "grad_norm": 0.6118027567863464,
      "learning_rate": 1.3775645653795881e-06,
      "loss": 0.4126,
      "num_input_tokens_seen": 22560536,
      "step": 34515
    },
    {
      "epoch": 18.09224318658281,
      "grad_norm": 0.4563140869140625,
      "learning_rate": 1.373823833652882e-06,
      "loss": 0.357,
      "num_input_tokens_seen": 22563832,
      "step": 34520
    },
    {
      "epoch": 18.094863731656183,
      "grad_norm": 0.4752199947834015,
      "learning_rate": 1.3700880442526876e-06,
      "loss": 0.4703,
      "num_input_tokens_seen": 22566616,
      "step": 34525
    },
    {
      "epoch": 18.09748427672956,
      "grad_norm": 0.2713492214679718,
      "learning_rate": 1.3663571979604884e-06,
      "loss": 0.4762,
      "num_input_tokens_seen": 22569976,
      "step": 34530
    },
    {
      "epoch": 18.100104821802937,
      "grad_norm": 0.5221849679946899,
      "learning_rate": 1.3626312955567406e-06,
      "loss": 0.5516,
      "num_input_tokens_seen": 22573112,
      "step": 34535
    },
    {
      "epoch": 18.10272536687631,
      "grad_norm": 0.5755932927131653,
      "learning_rate": 1.358910337820854e-06,
      "loss": 0.4375,
      "num_input_tokens_seen": 22576152,
      "step": 34540
    },
    {
      "epoch": 18.105345911949687,
      "grad_norm": 0.44207242131233215,
      "learning_rate": 1.3551943255312172e-06,
      "loss": 0.4595,
      "num_input_tokens_seen": 22579672,
      "step": 34545
    },
    {
      "epoch": 18.10796645702306,
      "grad_norm": 0.46734359860420227,
      "learning_rate": 1.3514832594651694e-06,
      "loss": 0.4222,
      "num_input_tokens_seen": 22583096,
      "step": 34550
    },
    {
      "epoch": 18.110587002096437,
      "grad_norm": 0.5221092104911804,
      "learning_rate": 1.347777140399026e-06,
      "loss": 0.5494,
      "num_input_tokens_seen": 22587256,
      "step": 34555
    },
    {
      "epoch": 18.11320754716981,
      "grad_norm": 0.5693758726119995,
      "learning_rate": 1.344075969108069e-06,
      "loss": 0.5654,
      "num_input_tokens_seen": 22591352,
      "step": 34560
    },
    {
      "epoch": 18.115828092243188,
      "grad_norm": 0.6701861619949341,
      "learning_rate": 1.3403797463665352e-06,
      "loss": 0.4331,
      "num_input_tokens_seen": 22594040,
      "step": 34565
    },
    {
      "epoch": 18.11844863731656,
      "grad_norm": 0.4679577052593231,
      "learning_rate": 1.3366884729476365e-06,
      "loss": 0.5241,
      "num_input_tokens_seen": 22597208,
      "step": 34570
    },
    {
      "epoch": 18.121069182389938,
      "grad_norm": 0.41770780086517334,
      "learning_rate": 1.3330021496235467e-06,
      "loss": 0.3157,
      "num_input_tokens_seen": 22600536,
      "step": 34575
    },
    {
      "epoch": 18.12368972746331,
      "grad_norm": 0.44973450899124146,
      "learning_rate": 1.3293207771653965e-06,
      "loss": 0.4871,
      "num_input_tokens_seen": 22603160,
      "step": 34580
    },
    {
      "epoch": 18.12631027253669,
      "grad_norm": 0.5015646815299988,
      "learning_rate": 1.3256443563432891e-06,
      "loss": 0.4031,
      "num_input_tokens_seen": 22606328,
      "step": 34585
    },
    {
      "epoch": 18.128930817610062,
      "grad_norm": 0.507161021232605,
      "learning_rate": 1.3219728879262956e-06,
      "loss": 0.4134,
      "num_input_tokens_seen": 22609624,
      "step": 34590
    },
    {
      "epoch": 18.13155136268344,
      "grad_norm": 0.5665279030799866,
      "learning_rate": 1.3183063726824375e-06,
      "loss": 0.5071,
      "num_input_tokens_seen": 22612728,
      "step": 34595
    },
    {
      "epoch": 18.134171907756812,
      "grad_norm": 0.38728171586990356,
      "learning_rate": 1.3146448113787151e-06,
      "loss": 0.498,
      "num_input_tokens_seen": 22615640,
      "step": 34600
    },
    {
      "epoch": 18.13679245283019,
      "grad_norm": 0.37220221757888794,
      "learning_rate": 1.3109882047810773e-06,
      "loss": 0.514,
      "num_input_tokens_seen": 22618360,
      "step": 34605
    },
    {
      "epoch": 18.139412997903563,
      "grad_norm": 0.47196805477142334,
      "learning_rate": 1.307336553654448e-06,
      "loss": 0.5444,
      "num_input_tokens_seen": 22621976,
      "step": 34610
    },
    {
      "epoch": 18.14203354297694,
      "grad_norm": 0.42027634382247925,
      "learning_rate": 1.303689858762716e-06,
      "loss": 0.5221,
      "num_input_tokens_seen": 22626008,
      "step": 34615
    },
    {
      "epoch": 18.144654088050313,
      "grad_norm": 0.5991038680076599,
      "learning_rate": 1.3000481208687188e-06,
      "loss": 0.3689,
      "num_input_tokens_seen": 22629368,
      "step": 34620
    },
    {
      "epoch": 18.14727463312369,
      "grad_norm": 0.3286205232143402,
      "learning_rate": 1.2964113407342743e-06,
      "loss": 0.3888,
      "num_input_tokens_seen": 22632440,
      "step": 34625
    },
    {
      "epoch": 18.149895178197063,
      "grad_norm": 0.6215473413467407,
      "learning_rate": 1.292779519120152e-06,
      "loss": 0.4167,
      "num_input_tokens_seen": 22635160,
      "step": 34630
    },
    {
      "epoch": 18.15251572327044,
      "grad_norm": 0.29775765538215637,
      "learning_rate": 1.2891526567860856e-06,
      "loss": 0.5197,
      "num_input_tokens_seen": 22639896,
      "step": 34635
    },
    {
      "epoch": 18.155136268343817,
      "grad_norm": 0.6642664670944214,
      "learning_rate": 1.2855307544907825e-06,
      "loss": 0.4184,
      "num_input_tokens_seen": 22642968,
      "step": 34640
    },
    {
      "epoch": 18.15775681341719,
      "grad_norm": 0.3931327164173126,
      "learning_rate": 1.281913812991889e-06,
      "loss": 0.3023,
      "num_input_tokens_seen": 22646456,
      "step": 34645
    },
    {
      "epoch": 18.160377358490567,
      "grad_norm": 0.3367750942707062,
      "learning_rate": 1.2783018330460389e-06,
      "loss": 0.4854,
      "num_input_tokens_seen": 22650648,
      "step": 34650
    },
    {
      "epoch": 18.16299790356394,
      "grad_norm": 0.44260871410369873,
      "learning_rate": 1.2746948154088111e-06,
      "loss": 0.3209,
      "num_input_tokens_seen": 22654296,
      "step": 34655
    },
    {
      "epoch": 18.165618448637318,
      "grad_norm": 0.42005404829978943,
      "learning_rate": 1.2710927608347605e-06,
      "loss": 0.325,
      "num_input_tokens_seen": 22658072,
      "step": 34660
    },
    {
      "epoch": 18.16823899371069,
      "grad_norm": 0.6788463592529297,
      "learning_rate": 1.2674956700773872e-06,
      "loss": 0.3291,
      "num_input_tokens_seen": 22661752,
      "step": 34665
    },
    {
      "epoch": 18.170859538784068,
      "grad_norm": 0.7252098321914673,
      "learning_rate": 1.2639035438891666e-06,
      "loss": 0.6437,
      "num_input_tokens_seen": 22665112,
      "step": 34670
    },
    {
      "epoch": 18.17348008385744,
      "grad_norm": 0.7329037189483643,
      "learning_rate": 1.260316383021526e-06,
      "loss": 0.3511,
      "num_input_tokens_seen": 22667640,
      "step": 34675
    },
    {
      "epoch": 18.17610062893082,
      "grad_norm": 0.4202706217765808,
      "learning_rate": 1.2567341882248645e-06,
      "loss": 0.4447,
      "num_input_tokens_seen": 22670840,
      "step": 34680
    },
    {
      "epoch": 18.178721174004192,
      "grad_norm": 0.5737042427062988,
      "learning_rate": 1.2531569602485305e-06,
      "loss": 0.4511,
      "num_input_tokens_seen": 22673720,
      "step": 34685
    },
    {
      "epoch": 18.18134171907757,
      "grad_norm": 0.6931536793708801,
      "learning_rate": 1.2495846998408445e-06,
      "loss": 0.4368,
      "num_input_tokens_seen": 22676984,
      "step": 34690
    },
    {
      "epoch": 18.183962264150942,
      "grad_norm": 0.34280574321746826,
      "learning_rate": 1.246017407749084e-06,
      "loss": 0.4116,
      "num_input_tokens_seen": 22680216,
      "step": 34695
    },
    {
      "epoch": 18.18658280922432,
      "grad_norm": 0.39273500442504883,
      "learning_rate": 1.2424550847194739e-06,
      "loss": 0.4278,
      "num_input_tokens_seen": 22683448,
      "step": 34700
    },
    {
      "epoch": 18.189203354297693,
      "grad_norm": 0.9542697668075562,
      "learning_rate": 1.2388977314972238e-06,
      "loss": 0.4179,
      "num_input_tokens_seen": 22689912,
      "step": 34705
    },
    {
      "epoch": 18.19182389937107,
      "grad_norm": 0.4066309928894043,
      "learning_rate": 1.235345348826486e-06,
      "loss": 0.2956,
      "num_input_tokens_seen": 22692696,
      "step": 34710
    },
    {
      "epoch": 18.194444444444443,
      "grad_norm": 0.4260020852088928,
      "learning_rate": 1.2317979374503825e-06,
      "loss": 0.4045,
      "num_input_tokens_seen": 22695032,
      "step": 34715
    },
    {
      "epoch": 18.19706498951782,
      "grad_norm": 0.5990619659423828,
      "learning_rate": 1.228255498110989e-06,
      "loss": 0.4299,
      "num_input_tokens_seen": 22698712,
      "step": 34720
    },
    {
      "epoch": 18.199685534591197,
      "grad_norm": 0.5828738808631897,
      "learning_rate": 1.224718031549338e-06,
      "loss": 0.5171,
      "num_input_tokens_seen": 22702200,
      "step": 34725
    },
    {
      "epoch": 18.20230607966457,
      "grad_norm": 0.536831259727478,
      "learning_rate": 1.2211855385054344e-06,
      "loss": 0.4105,
      "num_input_tokens_seen": 22704888,
      "step": 34730
    },
    {
      "epoch": 18.204926624737947,
      "grad_norm": 0.5199586153030396,
      "learning_rate": 1.217658019718232e-06,
      "loss": 0.3648,
      "num_input_tokens_seen": 22708280,
      "step": 34735
    },
    {
      "epoch": 18.20754716981132,
      "grad_norm": 1.4000117778778076,
      "learning_rate": 1.2141354759256484e-06,
      "loss": 0.5014,
      "num_input_tokens_seen": 22711384,
      "step": 34740
    },
    {
      "epoch": 18.210167714884697,
      "grad_norm": 0.258436381816864,
      "learning_rate": 1.2106179078645641e-06,
      "loss": 0.3853,
      "num_input_tokens_seen": 22714712,
      "step": 34745
    },
    {
      "epoch": 18.21278825995807,
      "grad_norm": 0.7119154334068298,
      "learning_rate": 1.2071053162708096e-06,
      "loss": 0.464,
      "num_input_tokens_seen": 22717432,
      "step": 34750
    },
    {
      "epoch": 18.215408805031448,
      "grad_norm": 0.47561559081077576,
      "learning_rate": 1.203597701879175e-06,
      "loss": 0.4347,
      "num_input_tokens_seen": 22720312,
      "step": 34755
    },
    {
      "epoch": 18.21802935010482,
      "grad_norm": 0.2681829035282135,
      "learning_rate": 1.2000950654234177e-06,
      "loss": 0.3598,
      "num_input_tokens_seen": 22727928,
      "step": 34760
    },
    {
      "epoch": 18.220649895178198,
      "grad_norm": 0.4095214307308197,
      "learning_rate": 1.1965974076362491e-06,
      "loss": 0.4062,
      "num_input_tokens_seen": 22730456,
      "step": 34765
    },
    {
      "epoch": 18.22327044025157,
      "grad_norm": 0.5967926979064941,
      "learning_rate": 1.193104729249342e-06,
      "loss": 0.4305,
      "num_input_tokens_seen": 22733240,
      "step": 34770
    },
    {
      "epoch": 18.22589098532495,
      "grad_norm": 0.3639604151248932,
      "learning_rate": 1.189617030993323e-06,
      "loss": 0.3744,
      "num_input_tokens_seen": 22736344,
      "step": 34775
    },
    {
      "epoch": 18.228511530398322,
      "grad_norm": 0.5669980049133301,
      "learning_rate": 1.1861343135977753e-06,
      "loss": 0.4563,
      "num_input_tokens_seen": 22739320,
      "step": 34780
    },
    {
      "epoch": 18.2311320754717,
      "grad_norm": 0.5622518658638,
      "learning_rate": 1.182656577791244e-06,
      "loss": 0.3788,
      "num_input_tokens_seen": 22743000,
      "step": 34785
    },
    {
      "epoch": 18.233752620545072,
      "grad_norm": 0.549935519695282,
      "learning_rate": 1.1791838243012359e-06,
      "loss": 0.4868,
      "num_input_tokens_seen": 22745752,
      "step": 34790
    },
    {
      "epoch": 18.23637316561845,
      "grad_norm": 0.5913494229316711,
      "learning_rate": 1.1757160538542117e-06,
      "loss": 0.5858,
      "num_input_tokens_seen": 22748504,
      "step": 34795
    },
    {
      "epoch": 18.238993710691823,
      "grad_norm": 0.4533625841140747,
      "learning_rate": 1.1722532671755882e-06,
      "loss": 0.4323,
      "num_input_tokens_seen": 22751704,
      "step": 34800
    },
    {
      "epoch": 18.2416142557652,
      "grad_norm": 0.48285943269729614,
      "learning_rate": 1.168795464989736e-06,
      "loss": 0.3979,
      "num_input_tokens_seen": 22755256,
      "step": 34805
    },
    {
      "epoch": 18.244234800838573,
      "grad_norm": 0.37961864471435547,
      "learning_rate": 1.165342648019993e-06,
      "loss": 0.6081,
      "num_input_tokens_seen": 22758616,
      "step": 34810
    },
    {
      "epoch": 18.24685534591195,
      "grad_norm": 0.7217199206352234,
      "learning_rate": 1.1618948169886457e-06,
      "loss": 0.3184,
      "num_input_tokens_seen": 22761432,
      "step": 34815
    },
    {
      "epoch": 18.249475890985323,
      "grad_norm": 0.36479344964027405,
      "learning_rate": 1.1584519726169446e-06,
      "loss": 0.5387,
      "num_input_tokens_seen": 22764376,
      "step": 34820
    },
    {
      "epoch": 18.2520964360587,
      "grad_norm": 0.5027823448181152,
      "learning_rate": 1.1550141156250942e-06,
      "loss": 0.4376,
      "num_input_tokens_seen": 22768056,
      "step": 34825
    },
    {
      "epoch": 18.254716981132077,
      "grad_norm": 0.4725695550441742,
      "learning_rate": 1.1515812467322556e-06,
      "loss": 0.5752,
      "num_input_tokens_seen": 22771128,
      "step": 34830
    },
    {
      "epoch": 18.25733752620545,
      "grad_norm": 0.6466461420059204,
      "learning_rate": 1.1481533666565375e-06,
      "loss": 0.3168,
      "num_input_tokens_seen": 22774360,
      "step": 34835
    },
    {
      "epoch": 18.259958071278827,
      "grad_norm": 0.36628878116607666,
      "learning_rate": 1.144730476115019e-06,
      "loss": 0.3513,
      "num_input_tokens_seen": 22778712,
      "step": 34840
    },
    {
      "epoch": 18.2625786163522,
      "grad_norm": 0.5376986265182495,
      "learning_rate": 1.1413125758237304e-06,
      "loss": 0.3962,
      "num_input_tokens_seen": 22781368,
      "step": 34845
    },
    {
      "epoch": 18.265199161425578,
      "grad_norm": 0.32186415791511536,
      "learning_rate": 1.1378996664976577e-06,
      "loss": 0.323,
      "num_input_tokens_seen": 22784120,
      "step": 34850
    },
    {
      "epoch": 18.26781970649895,
      "grad_norm": 0.38148677349090576,
      "learning_rate": 1.134491748850741e-06,
      "loss": 0.3318,
      "num_input_tokens_seen": 22787160,
      "step": 34855
    },
    {
      "epoch": 18.270440251572328,
      "grad_norm": 0.3251584470272064,
      "learning_rate": 1.13108882359588e-06,
      "loss": 0.4698,
      "num_input_tokens_seen": 22790584,
      "step": 34860
    },
    {
      "epoch": 18.2730607966457,
      "grad_norm": 0.36719033122062683,
      "learning_rate": 1.127690891444924e-06,
      "loss": 0.4401,
      "num_input_tokens_seen": 22794168,
      "step": 34865
    },
    {
      "epoch": 18.27568134171908,
      "grad_norm": 0.891727864742279,
      "learning_rate": 1.1242979531086829e-06,
      "loss": 0.4821,
      "num_input_tokens_seen": 22796600,
      "step": 34870
    },
    {
      "epoch": 18.278301886792452,
      "grad_norm": 0.5637738704681396,
      "learning_rate": 1.1209100092969244e-06,
      "loss": 0.5136,
      "num_input_tokens_seen": 22799896,
      "step": 34875
    },
    {
      "epoch": 18.28092243186583,
      "grad_norm": 0.4545198082923889,
      "learning_rate": 1.1175270607183625e-06,
      "loss": 0.5339,
      "num_input_tokens_seen": 22803416,
      "step": 34880
    },
    {
      "epoch": 18.283542976939202,
      "grad_norm": 0.48118075728416443,
      "learning_rate": 1.114149108080678e-06,
      "loss": 0.5192,
      "num_input_tokens_seen": 22806552,
      "step": 34885
    },
    {
      "epoch": 18.28616352201258,
      "grad_norm": 0.32364621758461,
      "learning_rate": 1.1107761520904925e-06,
      "loss": 0.4225,
      "num_input_tokens_seen": 22809496,
      "step": 34890
    },
    {
      "epoch": 18.288784067085953,
      "grad_norm": 0.3669945299625397,
      "learning_rate": 1.1074081934533908e-06,
      "loss": 0.3468,
      "num_input_tokens_seen": 22812664,
      "step": 34895
    },
    {
      "epoch": 18.29140461215933,
      "grad_norm": 0.4108451306819916,
      "learning_rate": 1.1040452328739153e-06,
      "loss": 0.3367,
      "num_input_tokens_seen": 22815992,
      "step": 34900
    },
    {
      "epoch": 18.294025157232703,
      "grad_norm": 0.3076004981994629,
      "learning_rate": 1.1006872710555615e-06,
      "loss": 0.4365,
      "num_input_tokens_seen": 22818968,
      "step": 34905
    },
    {
      "epoch": 18.29664570230608,
      "grad_norm": 0.5029816627502441,
      "learning_rate": 1.0973343087007703e-06,
      "loss": 0.3872,
      "num_input_tokens_seen": 22821624,
      "step": 34910
    },
    {
      "epoch": 18.299266247379457,
      "grad_norm": 0.6264298558235168,
      "learning_rate": 1.0939863465109473e-06,
      "loss": 0.5748,
      "num_input_tokens_seen": 22824696,
      "step": 34915
    },
    {
      "epoch": 18.30188679245283,
      "grad_norm": 0.4291306436061859,
      "learning_rate": 1.0906433851864433e-06,
      "loss": 0.4124,
      "num_input_tokens_seen": 22827864,
      "step": 34920
    },
    {
      "epoch": 18.304507337526207,
      "grad_norm": 0.47309160232543945,
      "learning_rate": 1.0873054254265713e-06,
      "loss": 0.6274,
      "num_input_tokens_seen": 22831384,
      "step": 34925
    },
    {
      "epoch": 18.30712788259958,
      "grad_norm": 0.46705806255340576,
      "learning_rate": 1.0839724679295976e-06,
      "loss": 0.5435,
      "num_input_tokens_seen": 22835928,
      "step": 34930
    },
    {
      "epoch": 18.309748427672957,
      "grad_norm": 0.39084237813949585,
      "learning_rate": 1.0806445133927311e-06,
      "loss": 0.4824,
      "num_input_tokens_seen": 22838776,
      "step": 34935
    },
    {
      "epoch": 18.31236897274633,
      "grad_norm": 0.45516300201416016,
      "learning_rate": 1.0773215625121457e-06,
      "loss": 0.4174,
      "num_input_tokens_seen": 22842008,
      "step": 34940
    },
    {
      "epoch": 18.314989517819708,
      "grad_norm": 0.6138992309570312,
      "learning_rate": 1.0740036159829715e-06,
      "loss": 0.5148,
      "num_input_tokens_seen": 22846040,
      "step": 34945
    },
    {
      "epoch": 18.31761006289308,
      "grad_norm": 0.6537005305290222,
      "learning_rate": 1.0706906744992724e-06,
      "loss": 0.3478,
      "num_input_tokens_seen": 22850328,
      "step": 34950
    },
    {
      "epoch": 18.320230607966458,
      "grad_norm": 0.8441603183746338,
      "learning_rate": 1.0673827387540857e-06,
      "loss": 0.5179,
      "num_input_tokens_seen": 22853624,
      "step": 34955
    },
    {
      "epoch": 18.32285115303983,
      "grad_norm": 0.8735643029212952,
      "learning_rate": 1.0640798094393916e-06,
      "loss": 0.6012,
      "num_input_tokens_seen": 22857240,
      "step": 34960
    },
    {
      "epoch": 18.32547169811321,
      "grad_norm": 0.5852388143539429,
      "learning_rate": 1.060781887246126e-06,
      "loss": 0.3901,
      "num_input_tokens_seen": 22860216,
      "step": 34965
    },
    {
      "epoch": 18.328092243186582,
      "grad_norm": 0.5904231667518616,
      "learning_rate": 1.057488972864179e-06,
      "loss": 0.3249,
      "num_input_tokens_seen": 22864920,
      "step": 34970
    },
    {
      "epoch": 18.33071278825996,
      "grad_norm": 0.5074658393859863,
      "learning_rate": 1.0542010669823855e-06,
      "loss": 0.4499,
      "num_input_tokens_seen": 22867928,
      "step": 34975
    },
    {
      "epoch": 18.333333333333332,
      "grad_norm": 0.4519503116607666,
      "learning_rate": 1.0509181702885395e-06,
      "loss": 0.4439,
      "num_input_tokens_seen": 22871256,
      "step": 34980
    },
    {
      "epoch": 18.33595387840671,
      "grad_norm": 1.0113563537597656,
      "learning_rate": 1.047640283469392e-06,
      "loss": 0.515,
      "num_input_tokens_seen": 22875544,
      "step": 34985
    },
    {
      "epoch": 18.338574423480082,
      "grad_norm": 0.47523245215415955,
      "learning_rate": 1.0443674072106308e-06,
      "loss": 0.489,
      "num_input_tokens_seen": 22878456,
      "step": 34990
    },
    {
      "epoch": 18.34119496855346,
      "grad_norm": 0.4111649692058563,
      "learning_rate": 1.041099542196905e-06,
      "loss": 0.5358,
      "num_input_tokens_seen": 22882168,
      "step": 34995
    },
    {
      "epoch": 18.343815513626833,
      "grad_norm": 0.45008766651153564,
      "learning_rate": 1.0378366891118234e-06,
      "loss": 0.3864,
      "num_input_tokens_seen": 22884824,
      "step": 35000
    },
    {
      "epoch": 18.34643605870021,
      "grad_norm": 0.4094920754432678,
      "learning_rate": 1.0345788486379265e-06,
      "loss": 0.4229,
      "num_input_tokens_seen": 22887864,
      "step": 35005
    },
    {
      "epoch": 18.349056603773583,
      "grad_norm": 0.44773346185684204,
      "learning_rate": 1.0313260214567272e-06,
      "loss": 0.5295,
      "num_input_tokens_seen": 22892184,
      "step": 35010
    },
    {
      "epoch": 18.35167714884696,
      "grad_norm": 0.5362904667854309,
      "learning_rate": 1.0280782082486729e-06,
      "loss": 0.4884,
      "num_input_tokens_seen": 22895096,
      "step": 35015
    },
    {
      "epoch": 18.354297693920337,
      "grad_norm": 0.4683229923248291,
      "learning_rate": 1.02483540969317e-06,
      "loss": 0.4502,
      "num_input_tokens_seen": 22898168,
      "step": 35020
    },
    {
      "epoch": 18.35691823899371,
      "grad_norm": 0.2473376840353012,
      "learning_rate": 1.0215976264685763e-06,
      "loss": 0.4242,
      "num_input_tokens_seen": 22901368,
      "step": 35025
    },
    {
      "epoch": 18.359538784067087,
      "grad_norm": 0.34691277146339417,
      "learning_rate": 1.0183648592522021e-06,
      "loss": 0.4983,
      "num_input_tokens_seen": 22905048,
      "step": 35030
    },
    {
      "epoch": 18.36215932914046,
      "grad_norm": 0.4698498249053955,
      "learning_rate": 1.0151371087203015e-06,
      "loss": 0.4387,
      "num_input_tokens_seen": 22907800,
      "step": 35035
    },
    {
      "epoch": 18.364779874213838,
      "grad_norm": 0.5182545185089111,
      "learning_rate": 1.0119143755480815e-06,
      "loss": 0.4694,
      "num_input_tokens_seen": 22911480,
      "step": 35040
    },
    {
      "epoch": 18.36740041928721,
      "grad_norm": 0.49801376461982727,
      "learning_rate": 1.0086966604097026e-06,
      "loss": 0.3514,
      "num_input_tokens_seen": 22914936,
      "step": 35045
    },
    {
      "epoch": 18.370020964360588,
      "grad_norm": 0.5065197944641113,
      "learning_rate": 1.0054839639782766e-06,
      "loss": 0.6722,
      "num_input_tokens_seen": 22918424,
      "step": 35050
    },
    {
      "epoch": 18.37264150943396,
      "grad_norm": 0.4445078372955322,
      "learning_rate": 1.00227628692586e-06,
      "loss": 0.4739,
      "num_input_tokens_seen": 22921112,
      "step": 35055
    },
    {
      "epoch": 18.37526205450734,
      "grad_norm": 0.5951824188232422,
      "learning_rate": 9.990736299234692e-07,
      "loss": 0.455,
      "num_input_tokens_seen": 22924024,
      "step": 35060
    },
    {
      "epoch": 18.377882599580712,
      "grad_norm": 0.6266577243804932,
      "learning_rate": 9.958759936410573e-07,
      "loss": 0.3361,
      "num_input_tokens_seen": 22927128,
      "step": 35065
    },
    {
      "epoch": 18.38050314465409,
      "grad_norm": 0.45267361402511597,
      "learning_rate": 9.926833787475304e-07,
      "loss": 0.2889,
      "num_input_tokens_seen": 22929912,
      "step": 35070
    },
    {
      "epoch": 18.383123689727462,
      "grad_norm": 0.42618054151535034,
      "learning_rate": 9.89495785910749e-07,
      "loss": 0.4138,
      "num_input_tokens_seen": 22932504,
      "step": 35075
    },
    {
      "epoch": 18.38574423480084,
      "grad_norm": 0.47301846742630005,
      "learning_rate": 9.86313215797524e-07,
      "loss": 0.3486,
      "num_input_tokens_seen": 22936600,
      "step": 35080
    },
    {
      "epoch": 18.388364779874212,
      "grad_norm": 0.9561731815338135,
      "learning_rate": 9.831356690736149e-07,
      "loss": 0.4339,
      "num_input_tokens_seen": 22939864,
      "step": 35085
    },
    {
      "epoch": 18.39098532494759,
      "grad_norm": 0.5487870573997498,
      "learning_rate": 9.799631464037256e-07,
      "loss": 0.5633,
      "num_input_tokens_seen": 22943480,
      "step": 35090
    },
    {
      "epoch": 18.393605870020963,
      "grad_norm": 0.44187524914741516,
      "learning_rate": 9.76795648451509e-07,
      "loss": 0.4976,
      "num_input_tokens_seen": 22946712,
      "step": 35095
    },
    {
      "epoch": 18.39622641509434,
      "grad_norm": 0.5602723956108093,
      "learning_rate": 9.73633175879568e-07,
      "loss": 0.4649,
      "num_input_tokens_seen": 22950392,
      "step": 35100
    },
    {
      "epoch": 18.398846960167713,
      "grad_norm": 0.3673449456691742,
      "learning_rate": 9.704757293494627e-07,
      "loss": 0.4628,
      "num_input_tokens_seen": 22954264,
      "step": 35105
    },
    {
      "epoch": 18.40146750524109,
      "grad_norm": 0.3678986132144928,
      "learning_rate": 9.673233095216922e-07,
      "loss": 0.3579,
      "num_input_tokens_seen": 22958232,
      "step": 35110
    },
    {
      "epoch": 18.404088050314467,
      "grad_norm": 0.8730214238166809,
      "learning_rate": 9.641759170557068e-07,
      "loss": 0.6253,
      "num_input_tokens_seen": 22961208,
      "step": 35115
    },
    {
      "epoch": 18.40670859538784,
      "grad_norm": 0.49882107973098755,
      "learning_rate": 9.610335526099024e-07,
      "loss": 0.3396,
      "num_input_tokens_seen": 22966296,
      "step": 35120
    },
    {
      "epoch": 18.409329140461217,
      "grad_norm": 0.5456153750419617,
      "learning_rate": 9.578962168416278e-07,
      "loss": 0.4084,
      "num_input_tokens_seen": 22968856,
      "step": 35125
    },
    {
      "epoch": 18.41194968553459,
      "grad_norm": 0.49980810284614563,
      "learning_rate": 9.547639104071775e-07,
      "loss": 0.291,
      "num_input_tokens_seen": 22972056,
      "step": 35130
    },
    {
      "epoch": 18.414570230607968,
      "grad_norm": 0.662371039390564,
      "learning_rate": 9.516366339617944e-07,
      "loss": 0.322,
      "num_input_tokens_seen": 22974840,
      "step": 35135
    },
    {
      "epoch": 18.41719077568134,
      "grad_norm": 0.4075905680656433,
      "learning_rate": 9.485143881596714e-07,
      "loss": 0.4855,
      "num_input_tokens_seen": 22978520,
      "step": 35140
    },
    {
      "epoch": 18.419811320754718,
      "grad_norm": 0.5391222834587097,
      "learning_rate": 9.453971736539447e-07,
      "loss": 0.4048,
      "num_input_tokens_seen": 22982136,
      "step": 35145
    },
    {
      "epoch": 18.42243186582809,
      "grad_norm": 0.8584864139556885,
      "learning_rate": 9.42284991096698e-07,
      "loss": 0.4343,
      "num_input_tokens_seen": 22985304,
      "step": 35150
    },
    {
      "epoch": 18.42505241090147,
      "grad_norm": 0.6299633383750916,
      "learning_rate": 9.391778411389634e-07,
      "loss": 0.404,
      "num_input_tokens_seen": 22988856,
      "step": 35155
    },
    {
      "epoch": 18.427672955974842,
      "grad_norm": 0.34617137908935547,
      "learning_rate": 9.360757244307267e-07,
      "loss": 0.3705,
      "num_input_tokens_seen": 22991896,
      "step": 35160
    },
    {
      "epoch": 18.43029350104822,
      "grad_norm": 0.4569626748561859,
      "learning_rate": 9.329786416209157e-07,
      "loss": 0.4201,
      "num_input_tokens_seen": 22994808,
      "step": 35165
    },
    {
      "epoch": 18.432914046121592,
      "grad_norm": 0.40693724155426025,
      "learning_rate": 9.298865933574014e-07,
      "loss": 0.4337,
      "num_input_tokens_seen": 22997560,
      "step": 35170
    },
    {
      "epoch": 18.43553459119497,
      "grad_norm": 0.3389033377170563,
      "learning_rate": 9.267995802870022e-07,
      "loss": 0.5634,
      "num_input_tokens_seen": 23000696,
      "step": 35175
    },
    {
      "epoch": 18.438155136268342,
      "grad_norm": 0.8408524990081787,
      "learning_rate": 9.237176030554878e-07,
      "loss": 0.4334,
      "num_input_tokens_seen": 23003320,
      "step": 35180
    },
    {
      "epoch": 18.44077568134172,
      "grad_norm": 0.6779108643531799,
      "learning_rate": 9.206406623075758e-07,
      "loss": 0.4526,
      "num_input_tokens_seen": 23006456,
      "step": 35185
    },
    {
      "epoch": 18.443396226415093,
      "grad_norm": 0.27354928851127625,
      "learning_rate": 9.175687586869292e-07,
      "loss": 0.4023,
      "num_input_tokens_seen": 23010712,
      "step": 35190
    },
    {
      "epoch": 18.44601677148847,
      "grad_norm": 0.5962302684783936,
      "learning_rate": 9.145018928361476e-07,
      "loss": 0.4416,
      "num_input_tokens_seen": 23014232,
      "step": 35195
    },
    {
      "epoch": 18.448637316561843,
      "grad_norm": 0.4591248035430908,
      "learning_rate": 9.114400653967903e-07,
      "loss": 0.3944,
      "num_input_tokens_seen": 23016952,
      "step": 35200
    },
    {
      "epoch": 18.45125786163522,
      "grad_norm": 0.4768452048301697,
      "learning_rate": 9.083832770093531e-07,
      "loss": 0.4174,
      "num_input_tokens_seen": 23019832,
      "step": 35205
    },
    {
      "epoch": 18.453878406708597,
      "grad_norm": 0.6519429683685303,
      "learning_rate": 9.053315283132802e-07,
      "loss": 0.5148,
      "num_input_tokens_seen": 23022264,
      "step": 35210
    },
    {
      "epoch": 18.45649895178197,
      "grad_norm": 0.6955143809318542,
      "learning_rate": 9.022848199469663e-07,
      "loss": 0.3709,
      "num_input_tokens_seen": 23025912,
      "step": 35215
    },
    {
      "epoch": 18.459119496855347,
      "grad_norm": 0.49635717272758484,
      "learning_rate": 8.992431525477518e-07,
      "loss": 0.3736,
      "num_input_tokens_seen": 23029560,
      "step": 35220
    },
    {
      "epoch": 18.46174004192872,
      "grad_norm": 0.5687522888183594,
      "learning_rate": 8.962065267519082e-07,
      "loss": 0.4491,
      "num_input_tokens_seen": 23033176,
      "step": 35225
    },
    {
      "epoch": 18.464360587002098,
      "grad_norm": 0.5442982316017151,
      "learning_rate": 8.931749431946718e-07,
      "loss": 0.42,
      "num_input_tokens_seen": 23036568,
      "step": 35230
    },
    {
      "epoch": 18.46698113207547,
      "grad_norm": 0.9496111273765564,
      "learning_rate": 8.901484025102103e-07,
      "loss": 0.4599,
      "num_input_tokens_seen": 23039448,
      "step": 35235
    },
    {
      "epoch": 18.469601677148848,
      "grad_norm": 0.6887589693069458,
      "learning_rate": 8.871269053316422e-07,
      "loss": 0.4676,
      "num_input_tokens_seen": 23043384,
      "step": 35240
    },
    {
      "epoch": 18.47222222222222,
      "grad_norm": 0.6669881939888,
      "learning_rate": 8.841104522910343e-07,
      "loss": 0.5446,
      "num_input_tokens_seen": 23047128,
      "step": 35245
    },
    {
      "epoch": 18.4748427672956,
      "grad_norm": 0.4100183844566345,
      "learning_rate": 8.8109904401939e-07,
      "loss": 0.4741,
      "num_input_tokens_seen": 23050072,
      "step": 35250
    },
    {
      "epoch": 18.47746331236897,
      "grad_norm": 0.6125227808952332,
      "learning_rate": 8.780926811466639e-07,
      "loss": 0.443,
      "num_input_tokens_seen": 23053688,
      "step": 35255
    },
    {
      "epoch": 18.48008385744235,
      "grad_norm": 0.43867403268814087,
      "learning_rate": 8.7509136430175e-07,
      "loss": 0.3662,
      "num_input_tokens_seen": 23056440,
      "step": 35260
    },
    {
      "epoch": 18.482704402515722,
      "grad_norm": 0.4436434507369995,
      "learning_rate": 8.720950941124877e-07,
      "loss": 0.4057,
      "num_input_tokens_seen": 23060120,
      "step": 35265
    },
    {
      "epoch": 18.4853249475891,
      "grad_norm": 0.5839512944221497,
      "learning_rate": 8.691038712056732e-07,
      "loss": 0.5112,
      "num_input_tokens_seen": 23063544,
      "step": 35270
    },
    {
      "epoch": 18.487945492662472,
      "grad_norm": 0.5752666592597961,
      "learning_rate": 8.66117696207025e-07,
      "loss": 0.3528,
      "num_input_tokens_seen": 23067096,
      "step": 35275
    },
    {
      "epoch": 18.49056603773585,
      "grad_norm": 0.3227338194847107,
      "learning_rate": 8.631365697412186e-07,
      "loss": 0.3492,
      "num_input_tokens_seen": 23070328,
      "step": 35280
    },
    {
      "epoch": 18.493186582809223,
      "grad_norm": 0.3084564507007599,
      "learning_rate": 8.601604924318802e-07,
      "loss": 0.5239,
      "num_input_tokens_seen": 23074648,
      "step": 35285
    },
    {
      "epoch": 18.4958071278826,
      "grad_norm": 0.3790278732776642,
      "learning_rate": 8.571894649015588e-07,
      "loss": 0.4109,
      "num_input_tokens_seen": 23077400,
      "step": 35290
    },
    {
      "epoch": 18.498427672955973,
      "grad_norm": 0.5695798397064209,
      "learning_rate": 8.542234877717686e-07,
      "loss": 0.5431,
      "num_input_tokens_seen": 23079992,
      "step": 35295
    },
    {
      "epoch": 18.50104821802935,
      "grad_norm": 0.5832088589668274,
      "learning_rate": 8.512625616629549e-07,
      "loss": 0.3805,
      "num_input_tokens_seen": 23082936,
      "step": 35300
    },
    {
      "epoch": 18.503668763102727,
      "grad_norm": 0.5704883933067322,
      "learning_rate": 8.483066871945083e-07,
      "loss": 0.4004,
      "num_input_tokens_seen": 23085528,
      "step": 35305
    },
    {
      "epoch": 18.5062893081761,
      "grad_norm": 0.9576751589775085,
      "learning_rate": 8.453558649847676e-07,
      "loss": 0.4417,
      "num_input_tokens_seen": 23089016,
      "step": 35310
    },
    {
      "epoch": 18.508909853249477,
      "grad_norm": 0.48143553733825684,
      "learning_rate": 8.424100956510112e-07,
      "loss": 0.4177,
      "num_input_tokens_seen": 23091512,
      "step": 35315
    },
    {
      "epoch": 18.51153039832285,
      "grad_norm": 0.8944731950759888,
      "learning_rate": 8.394693798094544e-07,
      "loss": 0.5942,
      "num_input_tokens_seen": 23095096,
      "step": 35320
    },
    {
      "epoch": 18.514150943396228,
      "grad_norm": 0.30487170815467834,
      "learning_rate": 8.365337180752691e-07,
      "loss": 0.4498,
      "num_input_tokens_seen": 23098904,
      "step": 35325
    },
    {
      "epoch": 18.5167714884696,
      "grad_norm": 0.5974147319793701,
      "learning_rate": 8.336031110625558e-07,
      "loss": 0.3625,
      "num_input_tokens_seen": 23101944,
      "step": 35330
    },
    {
      "epoch": 18.519392033542978,
      "grad_norm": 0.43580368161201477,
      "learning_rate": 8.306775593843657e-07,
      "loss": 0.3958,
      "num_input_tokens_seen": 23105816,
      "step": 35335
    },
    {
      "epoch": 18.52201257861635,
      "grad_norm": 0.46906715631484985,
      "learning_rate": 8.277570636526955e-07,
      "loss": 0.4166,
      "num_input_tokens_seen": 23109304,
      "step": 35340
    },
    {
      "epoch": 18.52463312368973,
      "grad_norm": 0.3466472327709198,
      "learning_rate": 8.24841624478473e-07,
      "loss": 0.3969,
      "num_input_tokens_seen": 23112984,
      "step": 35345
    },
    {
      "epoch": 18.5272536687631,
      "grad_norm": 0.543310284614563,
      "learning_rate": 8.219312424715825e-07,
      "loss": 0.3817,
      "num_input_tokens_seen": 23115832,
      "step": 35350
    },
    {
      "epoch": 18.52987421383648,
      "grad_norm": 0.42104488611221313,
      "learning_rate": 8.190259182408345e-07,
      "loss": 0.4423,
      "num_input_tokens_seen": 23118872,
      "step": 35355
    },
    {
      "epoch": 18.532494758909852,
      "grad_norm": 0.47917383909225464,
      "learning_rate": 8.161256523939925e-07,
      "loss": 0.5592,
      "num_input_tokens_seen": 23121880,
      "step": 35360
    },
    {
      "epoch": 18.53511530398323,
      "grad_norm": 0.5513731837272644,
      "learning_rate": 8.132304455377576e-07,
      "loss": 0.4454,
      "num_input_tokens_seen": 23125528,
      "step": 35365
    },
    {
      "epoch": 18.537735849056602,
      "grad_norm": 0.44541415572166443,
      "learning_rate": 8.103402982777813e-07,
      "loss": 0.5514,
      "num_input_tokens_seen": 23130360,
      "step": 35370
    },
    {
      "epoch": 18.54035639412998,
      "grad_norm": 0.7347441911697388,
      "learning_rate": 8.07455211218644e-07,
      "loss": 0.3709,
      "num_input_tokens_seen": 23133912,
      "step": 35375
    },
    {
      "epoch": 18.542976939203353,
      "grad_norm": 0.3565088212490082,
      "learning_rate": 8.045751849638738e-07,
      "loss": 0.4738,
      "num_input_tokens_seen": 23136440,
      "step": 35380
    },
    {
      "epoch": 18.54559748427673,
      "grad_norm": 0.4000701308250427,
      "learning_rate": 8.017002201159363e-07,
      "loss": 0.4491,
      "num_input_tokens_seen": 23139544,
      "step": 35385
    },
    {
      "epoch": 18.548218029350103,
      "grad_norm": 0.38417646288871765,
      "learning_rate": 7.988303172762418e-07,
      "loss": 0.3764,
      "num_input_tokens_seen": 23144408,
      "step": 35390
    },
    {
      "epoch": 18.55083857442348,
      "grad_norm": 0.5861843228340149,
      "learning_rate": 7.959654770451463e-07,
      "loss": 0.4851,
      "num_input_tokens_seen": 23147960,
      "step": 35395
    },
    {
      "epoch": 18.553459119496857,
      "grad_norm": 0.7693242430686951,
      "learning_rate": 7.9310570002194e-07,
      "loss": 0.3764,
      "num_input_tokens_seen": 23151096,
      "step": 35400
    },
    {
      "epoch": 18.55607966457023,
      "grad_norm": 0.42255526781082153,
      "learning_rate": 7.902509868048552e-07,
      "loss": 0.4444,
      "num_input_tokens_seen": 23154424,
      "step": 35405
    },
    {
      "epoch": 18.558700209643607,
      "grad_norm": 0.4308096766471863,
      "learning_rate": 7.874013379910616e-07,
      "loss": 0.3299,
      "num_input_tokens_seen": 23157048,
      "step": 35410
    },
    {
      "epoch": 18.56132075471698,
      "grad_norm": 0.429951012134552,
      "learning_rate": 7.845567541766769e-07,
      "loss": 0.3286,
      "num_input_tokens_seen": 23160792,
      "step": 35415
    },
    {
      "epoch": 18.563941299790358,
      "grad_norm": 0.3549305498600006,
      "learning_rate": 7.817172359567526e-07,
      "loss": 0.4432,
      "num_input_tokens_seen": 23164248,
      "step": 35420
    },
    {
      "epoch": 18.56656184486373,
      "grad_norm": 0.8045684099197388,
      "learning_rate": 7.788827839252888e-07,
      "loss": 0.5404,
      "num_input_tokens_seen": 23167832,
      "step": 35425
    },
    {
      "epoch": 18.569182389937108,
      "grad_norm": 0.2992423176765442,
      "learning_rate": 7.760533986752222e-07,
      "loss": 0.3623,
      "num_input_tokens_seen": 23172152,
      "step": 35430
    },
    {
      "epoch": 18.57180293501048,
      "grad_norm": 0.4605315923690796,
      "learning_rate": 7.732290807984155e-07,
      "loss": 0.3998,
      "num_input_tokens_seen": 23175000,
      "step": 35435
    },
    {
      "epoch": 18.57442348008386,
      "grad_norm": 0.8350948691368103,
      "learning_rate": 7.704098308856933e-07,
      "loss": 0.4527,
      "num_input_tokens_seen": 23177976,
      "step": 35440
    },
    {
      "epoch": 18.57704402515723,
      "grad_norm": 0.5044392943382263,
      "learning_rate": 7.675956495268088e-07,
      "loss": 0.4275,
      "num_input_tokens_seen": 23180792,
      "step": 35445
    },
    {
      "epoch": 18.57966457023061,
      "grad_norm": 0.4499744772911072,
      "learning_rate": 7.647865373104551e-07,
      "loss": 0.6399,
      "num_input_tokens_seen": 23184344,
      "step": 35450
    },
    {
      "epoch": 18.582285115303982,
      "grad_norm": 0.41768017411231995,
      "learning_rate": 7.619824948242732e-07,
      "loss": 0.4343,
      "num_input_tokens_seen": 23187640,
      "step": 35455
    },
    {
      "epoch": 18.58490566037736,
      "grad_norm": 0.5616315603256226,
      "learning_rate": 7.5918352265483e-07,
      "loss": 0.6428,
      "num_input_tokens_seen": 23191256,
      "step": 35460
    },
    {
      "epoch": 18.587526205450732,
      "grad_norm": 0.47022745013237,
      "learning_rate": 7.563896213876348e-07,
      "loss": 0.5097,
      "num_input_tokens_seen": 23193944,
      "step": 35465
    },
    {
      "epoch": 18.59014675052411,
      "grad_norm": 0.5807628035545349,
      "learning_rate": 7.53600791607148e-07,
      "loss": 0.4903,
      "num_input_tokens_seen": 23197464,
      "step": 35470
    },
    {
      "epoch": 18.592767295597483,
      "grad_norm": 0.5115822553634644,
      "learning_rate": 7.508170338967585e-07,
      "loss": 0.3728,
      "num_input_tokens_seen": 23200024,
      "step": 35475
    },
    {
      "epoch": 18.59538784067086,
      "grad_norm": 0.4699878990650177,
      "learning_rate": 7.480383488387976e-07,
      "loss": 0.4474,
      "num_input_tokens_seen": 23202968,
      "step": 35480
    },
    {
      "epoch": 18.598008385744233,
      "grad_norm": 0.35129913687705994,
      "learning_rate": 7.452647370145338e-07,
      "loss": 0.3856,
      "num_input_tokens_seen": 23206296,
      "step": 35485
    },
    {
      "epoch": 18.60062893081761,
      "grad_norm": 0.3991049528121948,
      "learning_rate": 7.424961990041696e-07,
      "loss": 0.4463,
      "num_input_tokens_seen": 23209304,
      "step": 35490
    },
    {
      "epoch": 18.603249475890987,
      "grad_norm": 0.4876171052455902,
      "learning_rate": 7.397327353868583e-07,
      "loss": 0.3514,
      "num_input_tokens_seen": 23213080,
      "step": 35495
    },
    {
      "epoch": 18.60587002096436,
      "grad_norm": 0.3903108537197113,
      "learning_rate": 7.369743467406848e-07,
      "loss": 0.4141,
      "num_input_tokens_seen": 23215928,
      "step": 35500
    },
    {
      "epoch": 18.608490566037737,
      "grad_norm": 0.37222111225128174,
      "learning_rate": 7.34221033642668e-07,
      "loss": 0.4635,
      "num_input_tokens_seen": 23219320,
      "step": 35505
    },
    {
      "epoch": 18.61111111111111,
      "grad_norm": 0.3342600166797638,
      "learning_rate": 7.314727966687751e-07,
      "loss": 0.3761,
      "num_input_tokens_seen": 23222296,
      "step": 35510
    },
    {
      "epoch": 18.613731656184488,
      "grad_norm": 0.5521852970123291,
      "learning_rate": 7.287296363939045e-07,
      "loss": 0.4589,
      "num_input_tokens_seen": 23224728,
      "step": 35515
    },
    {
      "epoch": 18.61635220125786,
      "grad_norm": 0.44238796830177307,
      "learning_rate": 7.259915533918888e-07,
      "loss": 0.3912,
      "num_input_tokens_seen": 23228248,
      "step": 35520
    },
    {
      "epoch": 18.618972746331238,
      "grad_norm": 0.5442391037940979,
      "learning_rate": 7.232585482355086e-07,
      "loss": 0.429,
      "num_input_tokens_seen": 23232600,
      "step": 35525
    },
    {
      "epoch": 18.62159329140461,
      "grad_norm": 0.8265125155448914,
      "learning_rate": 7.205306214964736e-07,
      "loss": 0.47,
      "num_input_tokens_seen": 23235384,
      "step": 35530
    },
    {
      "epoch": 18.62421383647799,
      "grad_norm": 0.4126282036304474,
      "learning_rate": 7.178077737454436e-07,
      "loss": 0.4189,
      "num_input_tokens_seen": 23238744,
      "step": 35535
    },
    {
      "epoch": 18.62683438155136,
      "grad_norm": 0.43307971954345703,
      "learning_rate": 7.150900055519993e-07,
      "loss": 0.4273,
      "num_input_tokens_seen": 23241720,
      "step": 35540
    },
    {
      "epoch": 18.62945492662474,
      "grad_norm": 0.47440409660339355,
      "learning_rate": 7.123773174846693e-07,
      "loss": 0.4652,
      "num_input_tokens_seen": 23244696,
      "step": 35545
    },
    {
      "epoch": 18.632075471698112,
      "grad_norm": 0.28401556611061096,
      "learning_rate": 7.096697101109134e-07,
      "loss": 0.3383,
      "num_input_tokens_seen": 23248536,
      "step": 35550
    },
    {
      "epoch": 18.63469601677149,
      "grad_norm": 0.578471839427948,
      "learning_rate": 7.069671839971371e-07,
      "loss": 0.4013,
      "num_input_tokens_seen": 23251352,
      "step": 35555
    },
    {
      "epoch": 18.637316561844862,
      "grad_norm": 0.3832319378852844,
      "learning_rate": 7.042697397086795e-07,
      "loss": 0.3663,
      "num_input_tokens_seen": 23254520,
      "step": 35560
    },
    {
      "epoch": 18.63993710691824,
      "grad_norm": 0.3569200038909912,
      "learning_rate": 7.015773778098089e-07,
      "loss": 0.5391,
      "num_input_tokens_seen": 23261752,
      "step": 35565
    },
    {
      "epoch": 18.642557651991613,
      "grad_norm": 0.6257449984550476,
      "learning_rate": 6.988900988637415e-07,
      "loss": 0.5191,
      "num_input_tokens_seen": 23265176,
      "step": 35570
    },
    {
      "epoch": 18.64517819706499,
      "grad_norm": 0.7080578804016113,
      "learning_rate": 6.962079034326218e-07,
      "loss": 0.4187,
      "num_input_tokens_seen": 23267800,
      "step": 35575
    },
    {
      "epoch": 18.647798742138363,
      "grad_norm": 0.705398678779602,
      "learning_rate": 6.935307920775346e-07,
      "loss": 0.3731,
      "num_input_tokens_seen": 23270488,
      "step": 35580
    },
    {
      "epoch": 18.65041928721174,
      "grad_norm": 0.37556254863739014,
      "learning_rate": 6.90858765358507e-07,
      "loss": 0.4677,
      "num_input_tokens_seen": 23274008,
      "step": 35585
    },
    {
      "epoch": 18.653039832285117,
      "grad_norm": 0.8167909979820251,
      "learning_rate": 6.881918238344915e-07,
      "loss": 0.2868,
      "num_input_tokens_seen": 23277240,
      "step": 35590
    },
    {
      "epoch": 18.65566037735849,
      "grad_norm": 0.7410842180252075,
      "learning_rate": 6.855299680633808e-07,
      "loss": 0.4857,
      "num_input_tokens_seen": 23279768,
      "step": 35595
    },
    {
      "epoch": 18.658280922431867,
      "grad_norm": 0.6322548389434814,
      "learning_rate": 6.828731986020104e-07,
      "loss": 0.5263,
      "num_input_tokens_seen": 23283128,
      "step": 35600
    },
    {
      "epoch": 18.66090146750524,
      "grad_norm": 0.5962609052658081,
      "learning_rate": 6.802215160061381e-07,
      "loss": 0.3977,
      "num_input_tokens_seen": 23286424,
      "step": 35605
    },
    {
      "epoch": 18.663522012578618,
      "grad_norm": 0.43906715512275696,
      "learning_rate": 6.775749208304704e-07,
      "loss": 0.5128,
      "num_input_tokens_seen": 23289880,
      "step": 35610
    },
    {
      "epoch": 18.66614255765199,
      "grad_norm": 0.47192713618278503,
      "learning_rate": 6.749334136286478e-07,
      "loss": 0.461,
      "num_input_tokens_seen": 23293528,
      "step": 35615
    },
    {
      "epoch": 18.668763102725368,
      "grad_norm": 0.6349743008613586,
      "learning_rate": 6.722969949532365e-07,
      "loss": 0.3178,
      "num_input_tokens_seen": 23296952,
      "step": 35620
    },
    {
      "epoch": 18.67138364779874,
      "grad_norm": 0.5727132558822632,
      "learning_rate": 6.69665665355751e-07,
      "loss": 0.41,
      "num_input_tokens_seen": 23299384,
      "step": 35625
    },
    {
      "epoch": 18.67400419287212,
      "grad_norm": 0.5041460990905762,
      "learning_rate": 6.670394253866314e-07,
      "loss": 0.4321,
      "num_input_tokens_seen": 23306232,
      "step": 35630
    },
    {
      "epoch": 18.67662473794549,
      "grad_norm": 0.7022462487220764,
      "learning_rate": 6.644182755952605e-07,
      "loss": 0.5526,
      "num_input_tokens_seen": 23309144,
      "step": 35635
    },
    {
      "epoch": 18.67924528301887,
      "grad_norm": 0.42328330874443054,
      "learning_rate": 6.618022165299526e-07,
      "loss": 0.3925,
      "num_input_tokens_seen": 23312472,
      "step": 35640
    },
    {
      "epoch": 18.681865828092242,
      "grad_norm": 0.6979321837425232,
      "learning_rate": 6.591912487379531e-07,
      "loss": 0.3604,
      "num_input_tokens_seen": 23315320,
      "step": 35645
    },
    {
      "epoch": 18.68448637316562,
      "grad_norm": 0.7800664901733398,
      "learning_rate": 6.565853727654503e-07,
      "loss": 0.4461,
      "num_input_tokens_seen": 23318360,
      "step": 35650
    },
    {
      "epoch": 18.687106918238992,
      "grad_norm": 0.9891555905342102,
      "learning_rate": 6.539845891575636e-07,
      "loss": 0.379,
      "num_input_tokens_seen": 23320824,
      "step": 35655
    },
    {
      "epoch": 18.68972746331237,
      "grad_norm": 0.5443264245986938,
      "learning_rate": 6.513888984583466e-07,
      "loss": 0.4874,
      "num_input_tokens_seen": 23324216,
      "step": 35660
    },
    {
      "epoch": 18.692348008385743,
      "grad_norm": 1.2511396408081055,
      "learning_rate": 6.487983012107873e-07,
      "loss": 0.47,
      "num_input_tokens_seen": 23326840,
      "step": 35665
    },
    {
      "epoch": 18.69496855345912,
      "grad_norm": 0.8639793395996094,
      "learning_rate": 6.462127979568133e-07,
      "loss": 0.4505,
      "num_input_tokens_seen": 23329464,
      "step": 35670
    },
    {
      "epoch": 18.697589098532493,
      "grad_norm": 0.6568685173988342,
      "learning_rate": 6.436323892372781e-07,
      "loss": 0.569,
      "num_input_tokens_seen": 23332280,
      "step": 35675
    },
    {
      "epoch": 18.70020964360587,
      "grad_norm": 0.4298940896987915,
      "learning_rate": 6.410570755919721e-07,
      "loss": 0.4423,
      "num_input_tokens_seen": 23335384,
      "step": 35680
    },
    {
      "epoch": 18.702830188679247,
      "grad_norm": 0.32977867126464844,
      "learning_rate": 6.384868575596281e-07,
      "loss": 0.3947,
      "num_input_tokens_seen": 23338776,
      "step": 35685
    },
    {
      "epoch": 18.70545073375262,
      "grad_norm": 0.7009100317955017,
      "learning_rate": 6.359217356779024e-07,
      "loss": 0.531,
      "num_input_tokens_seen": 23341624,
      "step": 35690
    },
    {
      "epoch": 18.708071278825997,
      "grad_norm": 0.37225690484046936,
      "learning_rate": 6.333617104833878e-07,
      "loss": 0.4448,
      "num_input_tokens_seen": 23345144,
      "step": 35695
    },
    {
      "epoch": 18.71069182389937,
      "grad_norm": 0.2735484540462494,
      "learning_rate": 6.308067825116142e-07,
      "loss": 0.3196,
      "num_input_tokens_seen": 23350328,
      "step": 35700
    },
    {
      "epoch": 18.713312368972748,
      "grad_norm": 0.9836333394050598,
      "learning_rate": 6.282569522970433e-07,
      "loss": 0.3581,
      "num_input_tokens_seen": 23353240,
      "step": 35705
    },
    {
      "epoch": 18.71593291404612,
      "grad_norm": 0.7027480006217957,
      "learning_rate": 6.257122203730731e-07,
      "loss": 0.5114,
      "num_input_tokens_seen": 23356120,
      "step": 35710
    },
    {
      "epoch": 18.718553459119498,
      "grad_norm": 0.5819352269172668,
      "learning_rate": 6.231725872720251e-07,
      "loss": 0.5598,
      "num_input_tokens_seen": 23359032,
      "step": 35715
    },
    {
      "epoch": 18.72117400419287,
      "grad_norm": 0.3948044776916504,
      "learning_rate": 6.206380535251716e-07,
      "loss": 0.4541,
      "num_input_tokens_seen": 23363064,
      "step": 35720
    },
    {
      "epoch": 18.72379454926625,
      "grad_norm": 0.6574004888534546,
      "learning_rate": 6.181086196626995e-07,
      "loss": 0.4833,
      "num_input_tokens_seen": 23365432,
      "step": 35725
    },
    {
      "epoch": 18.72641509433962,
      "grad_norm": 0.6287956833839417,
      "learning_rate": 6.155842862137384e-07,
      "loss": 0.4971,
      "num_input_tokens_seen": 23368120,
      "step": 35730
    },
    {
      "epoch": 18.729035639413,
      "grad_norm": 0.50677090883255,
      "learning_rate": 6.130650537063548e-07,
      "loss": 0.4761,
      "num_input_tokens_seen": 23371576,
      "step": 35735
    },
    {
      "epoch": 18.731656184486372,
      "grad_norm": 0.3952987790107727,
      "learning_rate": 6.105509226675438e-07,
      "loss": 0.4089,
      "num_input_tokens_seen": 23376216,
      "step": 35740
    },
    {
      "epoch": 18.73427672955975,
      "grad_norm": 0.4046517014503479,
      "learning_rate": 6.08041893623229e-07,
      "loss": 0.3677,
      "num_input_tokens_seen": 23379256,
      "step": 35745
    },
    {
      "epoch": 18.736897274633122,
      "grad_norm": 0.8450313210487366,
      "learning_rate": 6.055379670982713e-07,
      "loss": 0.4182,
      "num_input_tokens_seen": 23381912,
      "step": 35750
    },
    {
      "epoch": 18.7395178197065,
      "grad_norm": 0.5671306848526001,
      "learning_rate": 6.030391436164628e-07,
      "loss": 0.4105,
      "num_input_tokens_seen": 23384568,
      "step": 35755
    },
    {
      "epoch": 18.742138364779873,
      "grad_norm": 0.4885002374649048,
      "learning_rate": 6.005454237005298e-07,
      "loss": 0.4258,
      "num_input_tokens_seen": 23387320,
      "step": 35760
    },
    {
      "epoch": 18.74475890985325,
      "grad_norm": 0.46546119451522827,
      "learning_rate": 5.980568078721327e-07,
      "loss": 0.3795,
      "num_input_tokens_seen": 23390168,
      "step": 35765
    },
    {
      "epoch": 18.747379454926623,
      "grad_norm": 0.535459578037262,
      "learning_rate": 5.955732966518606e-07,
      "loss": 0.4766,
      "num_input_tokens_seen": 23393080,
      "step": 35770
    },
    {
      "epoch": 18.75,
      "grad_norm": 0.4162834584712982,
      "learning_rate": 5.930948905592343e-07,
      "loss": 0.3252,
      "num_input_tokens_seen": 23395800,
      "step": 35775
    },
    {
      "epoch": 18.752620545073377,
      "grad_norm": 0.6040893197059631,
      "learning_rate": 5.906215901127054e-07,
      "loss": 0.3737,
      "num_input_tokens_seen": 23398424,
      "step": 35780
    },
    {
      "epoch": 18.75524109014675,
      "grad_norm": 1.0026851892471313,
      "learning_rate": 5.881533958296631e-07,
      "loss": 0.4337,
      "num_input_tokens_seen": 23400984,
      "step": 35785
    },
    {
      "epoch": 18.757861635220127,
      "grad_norm": 0.5042615532875061,
      "learning_rate": 5.856903082264248e-07,
      "loss": 0.4101,
      "num_input_tokens_seen": 23404952,
      "step": 35790
    },
    {
      "epoch": 18.7604821802935,
      "grad_norm": 0.47326797246932983,
      "learning_rate": 5.832323278182422e-07,
      "loss": 0.4625,
      "num_input_tokens_seen": 23409048,
      "step": 35795
    },
    {
      "epoch": 18.763102725366878,
      "grad_norm": 0.39125025272369385,
      "learning_rate": 5.807794551192986e-07,
      "loss": 0.4809,
      "num_input_tokens_seen": 23411864,
      "step": 35800
    },
    {
      "epoch": 18.76572327044025,
      "grad_norm": 0.6939232349395752,
      "learning_rate": 5.783316906427e-07,
      "loss": 0.3775,
      "num_input_tokens_seen": 23414936,
      "step": 35805
    },
    {
      "epoch": 18.768343815513628,
      "grad_norm": 0.5487970113754272,
      "learning_rate": 5.758890349004925e-07,
      "loss": 0.4494,
      "num_input_tokens_seen": 23417976,
      "step": 35810
    },
    {
      "epoch": 18.770964360587,
      "grad_norm": 0.4926478862762451,
      "learning_rate": 5.734514884036563e-07,
      "loss": 0.3707,
      "num_input_tokens_seen": 23420984,
      "step": 35815
    },
    {
      "epoch": 18.77358490566038,
      "grad_norm": 0.39952459931373596,
      "learning_rate": 5.710190516620945e-07,
      "loss": 0.3634,
      "num_input_tokens_seen": 23424120,
      "step": 35820
    },
    {
      "epoch": 18.77620545073375,
      "grad_norm": 0.47125867009162903,
      "learning_rate": 5.685917251846501e-07,
      "loss": 0.5033,
      "num_input_tokens_seen": 23426584,
      "step": 35825
    },
    {
      "epoch": 18.77882599580713,
      "grad_norm": 0.8610146641731262,
      "learning_rate": 5.661695094790892e-07,
      "loss": 0.4183,
      "num_input_tokens_seen": 23429880,
      "step": 35830
    },
    {
      "epoch": 18.781446540880502,
      "grad_norm": 0.6176340579986572,
      "learning_rate": 5.637524050521064e-07,
      "loss": 0.3819,
      "num_input_tokens_seen": 23432728,
      "step": 35835
    },
    {
      "epoch": 18.78406708595388,
      "grad_norm": 0.6240939497947693,
      "learning_rate": 5.613404124093391e-07,
      "loss": 0.5207,
      "num_input_tokens_seen": 23435704,
      "step": 35840
    },
    {
      "epoch": 18.786687631027252,
      "grad_norm": 0.9792578816413879,
      "learning_rate": 5.589335320553446e-07,
      "loss": 0.363,
      "num_input_tokens_seen": 23439416,
      "step": 35845
    },
    {
      "epoch": 18.78930817610063,
      "grad_norm": 0.5668495297431946,
      "learning_rate": 5.565317644936202e-07,
      "loss": 0.4427,
      "num_input_tokens_seen": 23441880,
      "step": 35850
    },
    {
      "epoch": 18.791928721174003,
      "grad_norm": 0.5574182868003845,
      "learning_rate": 5.541351102265863e-07,
      "loss": 0.4703,
      "num_input_tokens_seen": 23445560,
      "step": 35855
    },
    {
      "epoch": 18.79454926624738,
      "grad_norm": 0.5309727191925049,
      "learning_rate": 5.51743569755589e-07,
      "loss": 0.4889,
      "num_input_tokens_seen": 23448760,
      "step": 35860
    },
    {
      "epoch": 18.797169811320753,
      "grad_norm": 0.633850634098053,
      "learning_rate": 5.493571435809197e-07,
      "loss": 0.2792,
      "num_input_tokens_seen": 23451864,
      "step": 35865
    },
    {
      "epoch": 18.79979035639413,
      "grad_norm": 0.4341908097267151,
      "learning_rate": 5.469758322017849e-07,
      "loss": 0.3803,
      "num_input_tokens_seen": 23458552,
      "step": 35870
    },
    {
      "epoch": 18.802410901467507,
      "grad_norm": 0.5259716510772705,
      "learning_rate": 5.445996361163358e-07,
      "loss": 0.3883,
      "num_input_tokens_seen": 23461656,
      "step": 35875
    },
    {
      "epoch": 18.80503144654088,
      "grad_norm": 0.30730152130126953,
      "learning_rate": 5.422285558216389e-07,
      "loss": 0.4829,
      "num_input_tokens_seen": 23464664,
      "step": 35880
    },
    {
      "epoch": 18.807651991614257,
      "grad_norm": 0.5826876163482666,
      "learning_rate": 5.398625918137001e-07,
      "loss": 0.4863,
      "num_input_tokens_seen": 23468216,
      "step": 35885
    },
    {
      "epoch": 18.81027253668763,
      "grad_norm": 0.5262902975082397,
      "learning_rate": 5.375017445874486e-07,
      "loss": 0.4981,
      "num_input_tokens_seen": 23471064,
      "step": 35890
    },
    {
      "epoch": 18.812893081761008,
      "grad_norm": 0.3226843774318695,
      "learning_rate": 5.351460146367504e-07,
      "loss": 0.3526,
      "num_input_tokens_seen": 23474264,
      "step": 35895
    },
    {
      "epoch": 18.81551362683438,
      "grad_norm": 1.2754875421524048,
      "learning_rate": 5.327954024543946e-07,
      "loss": 0.3693,
      "num_input_tokens_seen": 23477720,
      "step": 35900
    },
    {
      "epoch": 18.818134171907758,
      "grad_norm": 0.5064606666564941,
      "learning_rate": 5.304499085321074e-07,
      "loss": 0.3315,
      "num_input_tokens_seen": 23480344,
      "step": 35905
    },
    {
      "epoch": 18.82075471698113,
      "grad_norm": 0.5031932592391968,
      "learning_rate": 5.281095333605352e-07,
      "loss": 0.4523,
      "num_input_tokens_seen": 23483544,
      "step": 35910
    },
    {
      "epoch": 18.82337526205451,
      "grad_norm": 0.36402827501296997,
      "learning_rate": 5.257742774292556e-07,
      "loss": 0.4548,
      "num_input_tokens_seen": 23486392,
      "step": 35915
    },
    {
      "epoch": 18.82599580712788,
      "grad_norm": 0.5256338119506836,
      "learning_rate": 5.234441412267837e-07,
      "loss": 0.6145,
      "num_input_tokens_seen": 23490200,
      "step": 35920
    },
    {
      "epoch": 18.82861635220126,
      "grad_norm": 0.25183817744255066,
      "learning_rate": 5.211191252405517e-07,
      "loss": 0.4177,
      "num_input_tokens_seen": 23493048,
      "step": 35925
    },
    {
      "epoch": 18.831236897274632,
      "grad_norm": 0.6139267683029175,
      "learning_rate": 5.187992299569316e-07,
      "loss": 0.5675,
      "num_input_tokens_seen": 23496440,
      "step": 35930
    },
    {
      "epoch": 18.83385744234801,
      "grad_norm": 0.3380448520183563,
      "learning_rate": 5.164844558612131e-07,
      "loss": 0.3977,
      "num_input_tokens_seen": 23499832,
      "step": 35935
    },
    {
      "epoch": 18.836477987421382,
      "grad_norm": 0.4754042327404022,
      "learning_rate": 5.141748034376254e-07,
      "loss": 0.517,
      "num_input_tokens_seen": 23503416,
      "step": 35940
    },
    {
      "epoch": 18.83909853249476,
      "grad_norm": 0.7373019456863403,
      "learning_rate": 5.118702731693209e-07,
      "loss": 0.3478,
      "num_input_tokens_seen": 23506616,
      "step": 35945
    },
    {
      "epoch": 18.841719077568133,
      "grad_norm": 0.4740626811981201,
      "learning_rate": 5.095708655383751e-07,
      "loss": 0.5474,
      "num_input_tokens_seen": 23509912,
      "step": 35950
    },
    {
      "epoch": 18.84433962264151,
      "grad_norm": 0.2801676392555237,
      "learning_rate": 5.07276581025809e-07,
      "loss": 0.4052,
      "num_input_tokens_seen": 23513272,
      "step": 35955
    },
    {
      "epoch": 18.846960167714883,
      "grad_norm": 0.4908902049064636,
      "learning_rate": 5.049874201115495e-07,
      "loss": 0.4136,
      "num_input_tokens_seen": 23515992,
      "step": 35960
    },
    {
      "epoch": 18.84958071278826,
      "grad_norm": 0.4981483817100525,
      "learning_rate": 5.027033832744693e-07,
      "loss": 0.394,
      "num_input_tokens_seen": 23519032,
      "step": 35965
    },
    {
      "epoch": 18.852201257861637,
      "grad_norm": 0.4142865836620331,
      "learning_rate": 5.004244709923639e-07,
      "loss": 0.3832,
      "num_input_tokens_seen": 23522392,
      "step": 35970
    },
    {
      "epoch": 18.85482180293501,
      "grad_norm": 0.46075865626335144,
      "learning_rate": 4.981506837419519e-07,
      "loss": 0.3428,
      "num_input_tokens_seen": 23525240,
      "step": 35975
    },
    {
      "epoch": 18.857442348008387,
      "grad_norm": 0.4597597122192383,
      "learning_rate": 4.958820219988836e-07,
      "loss": 0.4877,
      "num_input_tokens_seen": 23528568,
      "step": 35980
    },
    {
      "epoch": 18.86006289308176,
      "grad_norm": 0.5892096161842346,
      "learning_rate": 4.936184862377402e-07,
      "loss": 0.4265,
      "num_input_tokens_seen": 23531288,
      "step": 35985
    },
    {
      "epoch": 18.862683438155138,
      "grad_norm": 0.5044569373130798,
      "learning_rate": 4.913600769320237e-07,
      "loss": 0.4245,
      "num_input_tokens_seen": 23534520,
      "step": 35990
    },
    {
      "epoch": 18.86530398322851,
      "grad_norm": 0.4347462058067322,
      "learning_rate": 4.891067945541728e-07,
      "loss": 0.4887,
      "num_input_tokens_seen": 23537880,
      "step": 35995
    },
    {
      "epoch": 18.867924528301888,
      "grad_norm": 0.2670508027076721,
      "learning_rate": 4.868586395755437e-07,
      "loss": 0.4179,
      "num_input_tokens_seen": 23541464,
      "step": 36000
    },
    {
      "epoch": 18.87054507337526,
      "grad_norm": 0.7276753187179565,
      "learning_rate": 4.846156124664242e-07,
      "loss": 0.2769,
      "num_input_tokens_seen": 23544760,
      "step": 36005
    },
    {
      "epoch": 18.87316561844864,
      "grad_norm": 0.47487157583236694,
      "learning_rate": 4.823777136960362e-07,
      "loss": 0.4254,
      "num_input_tokens_seen": 23547832,
      "step": 36010
    },
    {
      "epoch": 18.87578616352201,
      "grad_norm": 0.4658547639846802,
      "learning_rate": 4.801449437325162e-07,
      "loss": 0.3465,
      "num_input_tokens_seen": 23550968,
      "step": 36015
    },
    {
      "epoch": 18.87840670859539,
      "grad_norm": 0.48802649974823,
      "learning_rate": 4.779173030429379e-07,
      "loss": 0.3724,
      "num_input_tokens_seen": 23554680,
      "step": 36020
    },
    {
      "epoch": 18.881027253668762,
      "grad_norm": 0.3305223286151886,
      "learning_rate": 4.7569479209329804e-07,
      "loss": 0.4109,
      "num_input_tokens_seen": 23558168,
      "step": 36025
    },
    {
      "epoch": 18.88364779874214,
      "grad_norm": 0.4968716502189636,
      "learning_rate": 4.734774113485163e-07,
      "loss": 0.4269,
      "num_input_tokens_seen": 23560952,
      "step": 36030
    },
    {
      "epoch": 18.886268343815512,
      "grad_norm": 0.30417782068252563,
      "learning_rate": 4.7126516127244945e-07,
      "loss": 0.3863,
      "num_input_tokens_seen": 23563096,
      "step": 36035
    },
    {
      "epoch": 18.88888888888889,
      "grad_norm": 0.4405229687690735,
      "learning_rate": 4.690580423278718e-07,
      "loss": 0.5251,
      "num_input_tokens_seen": 23565400,
      "step": 36040
    },
    {
      "epoch": 18.891509433962263,
      "grad_norm": 0.5248746871948242,
      "learning_rate": 4.668560549764861e-07,
      "loss": 0.3989,
      "num_input_tokens_seen": 23569016,
      "step": 36045
    },
    {
      "epoch": 18.89412997903564,
      "grad_norm": 0.4490797221660614,
      "learning_rate": 4.646591996789268e-07,
      "loss": 0.486,
      "num_input_tokens_seen": 23572216,
      "step": 36050
    },
    {
      "epoch": 18.896750524109013,
      "grad_norm": 0.33313924074172974,
      "learning_rate": 4.6246747689474847e-07,
      "loss": 0.498,
      "num_input_tokens_seen": 23576056,
      "step": 36055
    },
    {
      "epoch": 18.89937106918239,
      "grad_norm": 0.2501473128795624,
      "learning_rate": 4.6028088708243434e-07,
      "loss": 0.571,
      "num_input_tokens_seen": 23579096,
      "step": 36060
    },
    {
      "epoch": 18.901991614255767,
      "grad_norm": 0.7077343463897705,
      "learning_rate": 4.580994306993991e-07,
      "loss": 0.3949,
      "num_input_tokens_seen": 23581720,
      "step": 36065
    },
    {
      "epoch": 18.90461215932914,
      "grad_norm": 0.748441755771637,
      "learning_rate": 4.5592310820196936e-07,
      "loss": 0.6488,
      "num_input_tokens_seen": 23584824,
      "step": 36070
    },
    {
      "epoch": 18.907232704402517,
      "grad_norm": 0.9201707243919373,
      "learning_rate": 4.537519200454116e-07,
      "loss": 0.4302,
      "num_input_tokens_seen": 23587800,
      "step": 36075
    },
    {
      "epoch": 18.90985324947589,
      "grad_norm": 0.4338831603527069,
      "learning_rate": 4.515858666839151e-07,
      "loss": 0.5668,
      "num_input_tokens_seen": 23590584,
      "step": 36080
    },
    {
      "epoch": 18.912473794549268,
      "grad_norm": 0.39228808879852295,
      "learning_rate": 4.494249485705926e-07,
      "loss": 0.6061,
      "num_input_tokens_seen": 23594392,
      "step": 36085
    },
    {
      "epoch": 18.91509433962264,
      "grad_norm": 0.38742512464523315,
      "learning_rate": 4.472691661574824e-07,
      "loss": 0.404,
      "num_input_tokens_seen": 23597176,
      "step": 36090
    },
    {
      "epoch": 18.917714884696018,
      "grad_norm": 0.4335341155529022,
      "learning_rate": 4.4511851989555156e-07,
      "loss": 0.4454,
      "num_input_tokens_seen": 23600600,
      "step": 36095
    },
    {
      "epoch": 18.92033542976939,
      "grad_norm": 0.498473197221756,
      "learning_rate": 4.429730102346874e-07,
      "loss": 0.5492,
      "num_input_tokens_seen": 23603480,
      "step": 36100
    },
    {
      "epoch": 18.92295597484277,
      "grad_norm": 0.48254701495170593,
      "learning_rate": 4.4083263762370873e-07,
      "loss": 0.3487,
      "num_input_tokens_seen": 23605784,
      "step": 36105
    },
    {
      "epoch": 18.92557651991614,
      "grad_norm": 0.44586852192878723,
      "learning_rate": 4.386974025103574e-07,
      "loss": 0.4229,
      "num_input_tokens_seen": 23609144,
      "step": 36110
    },
    {
      "epoch": 18.92819706498952,
      "grad_norm": 0.5061739087104797,
      "learning_rate": 4.365673053413011e-07,
      "loss": 0.5055,
      "num_input_tokens_seen": 23611960,
      "step": 36115
    },
    {
      "epoch": 18.930817610062892,
      "grad_norm": 0.5284525156021118,
      "learning_rate": 4.34442346562125e-07,
      "loss": 0.4323,
      "num_input_tokens_seen": 23614680,
      "step": 36120
    },
    {
      "epoch": 18.93343815513627,
      "grad_norm": 0.5834205150604248,
      "learning_rate": 4.323225266173542e-07,
      "loss": 0.4111,
      "num_input_tokens_seen": 23617592,
      "step": 36125
    },
    {
      "epoch": 18.936058700209642,
      "grad_norm": 1.1981606483459473,
      "learning_rate": 4.302078459504283e-07,
      "loss": 0.431,
      "num_input_tokens_seen": 23620440,
      "step": 36130
    },
    {
      "epoch": 18.93867924528302,
      "grad_norm": 0.2886433005332947,
      "learning_rate": 4.2809830500371294e-07,
      "loss": 0.3357,
      "num_input_tokens_seen": 23623320,
      "step": 36135
    },
    {
      "epoch": 18.941299790356393,
      "grad_norm": 0.6982663869857788,
      "learning_rate": 4.2599390421850236e-07,
      "loss": 0.3469,
      "num_input_tokens_seen": 23626168,
      "step": 36140
    },
    {
      "epoch": 18.94392033542977,
      "grad_norm": 0.6106567978858948,
      "learning_rate": 4.2389464403501113e-07,
      "loss": 0.5352,
      "num_input_tokens_seen": 23629240,
      "step": 36145
    },
    {
      "epoch": 18.946540880503143,
      "grad_norm": 0.4605480134487152,
      "learning_rate": 4.218005248923823e-07,
      "loss": 0.4056,
      "num_input_tokens_seen": 23632472,
      "step": 36150
    },
    {
      "epoch": 18.94916142557652,
      "grad_norm": 0.4107097089290619,
      "learning_rate": 4.1971154722867946e-07,
      "loss": 0.4789,
      "num_input_tokens_seen": 23635864,
      "step": 36155
    },
    {
      "epoch": 18.951781970649897,
      "grad_norm": 0.4649865925312042,
      "learning_rate": 4.176277114808946e-07,
      "loss": 0.4059,
      "num_input_tokens_seen": 23638712,
      "step": 36160
    },
    {
      "epoch": 18.95440251572327,
      "grad_norm": 0.6139723062515259,
      "learning_rate": 4.155490180849431e-07,
      "loss": 0.4272,
      "num_input_tokens_seen": 23641592,
      "step": 36165
    },
    {
      "epoch": 18.957023060796647,
      "grad_norm": 0.35522791743278503,
      "learning_rate": 4.1347546747566304e-07,
      "loss": 0.404,
      "num_input_tokens_seen": 23645912,
      "step": 36170
    },
    {
      "epoch": 18.95964360587002,
      "grad_norm": 0.36059147119522095,
      "learning_rate": 4.114070600868131e-07,
      "loss": 0.4262,
      "num_input_tokens_seen": 23648824,
      "step": 36175
    },
    {
      "epoch": 18.962264150943398,
      "grad_norm": 0.6487425565719604,
      "learning_rate": 4.093437963510888e-07,
      "loss": 0.4309,
      "num_input_tokens_seen": 23651096,
      "step": 36180
    },
    {
      "epoch": 18.96488469601677,
      "grad_norm": 0.6209201812744141,
      "learning_rate": 4.072856767000949e-07,
      "loss": 0.4364,
      "num_input_tokens_seen": 23654616,
      "step": 36185
    },
    {
      "epoch": 18.967505241090148,
      "grad_norm": 0.3170095682144165,
      "learning_rate": 4.052327015643703e-07,
      "loss": 0.4896,
      "num_input_tokens_seen": 23658200,
      "step": 36190
    },
    {
      "epoch": 18.97012578616352,
      "grad_norm": 0.692929744720459,
      "learning_rate": 4.031848713733744e-07,
      "loss": 0.4705,
      "num_input_tokens_seen": 23660856,
      "step": 36195
    },
    {
      "epoch": 18.9727463312369,
      "grad_norm": 0.3735884428024292,
      "learning_rate": 4.011421865554837e-07,
      "loss": 0.4389,
      "num_input_tokens_seen": 23664056,
      "step": 36200
    },
    {
      "epoch": 18.97536687631027,
      "grad_norm": 0.3741215467453003,
      "learning_rate": 3.9910464753801224e-07,
      "loss": 0.4751,
      "num_input_tokens_seen": 23667256,
      "step": 36205
    },
    {
      "epoch": 18.97798742138365,
      "grad_norm": 0.6103509664535522,
      "learning_rate": 3.9707225474718557e-07,
      "loss": 0.465,
      "num_input_tokens_seen": 23670456,
      "step": 36210
    },
    {
      "epoch": 18.980607966457022,
      "grad_norm": 1.3241255283355713,
      "learning_rate": 3.950450086081553e-07,
      "loss": 0.4349,
      "num_input_tokens_seen": 23673656,
      "step": 36215
    },
    {
      "epoch": 18.9832285115304,
      "grad_norm": 0.5926249027252197,
      "learning_rate": 3.930229095450072e-07,
      "loss": 0.503,
      "num_input_tokens_seen": 23676824,
      "step": 36220
    },
    {
      "epoch": 18.985849056603772,
      "grad_norm": 0.5947101712226868,
      "learning_rate": 3.9100595798073345e-07,
      "loss": 0.6729,
      "num_input_tokens_seen": 23679832,
      "step": 36225
    },
    {
      "epoch": 18.98846960167715,
      "grad_norm": 0.8098554015159607,
      "learning_rate": 3.8899415433725493e-07,
      "loss": 0.415,
      "num_input_tokens_seen": 23682808,
      "step": 36230
    },
    {
      "epoch": 18.991090146750523,
      "grad_norm": 0.7203764915466309,
      "learning_rate": 3.8698749903542666e-07,
      "loss": 0.5373,
      "num_input_tokens_seen": 23685976,
      "step": 36235
    },
    {
      "epoch": 18.9937106918239,
      "grad_norm": 0.8267771005630493,
      "learning_rate": 3.8498599249501e-07,
      "loss": 0.499,
      "num_input_tokens_seen": 23688952,
      "step": 36240
    },
    {
      "epoch": 18.996331236897273,
      "grad_norm": 0.6652708649635315,
      "learning_rate": 3.8298963513470345e-07,
      "loss": 0.4406,
      "num_input_tokens_seen": 23691832,
      "step": 36245
    },
    {
      "epoch": 18.99895178197065,
      "grad_norm": 0.3840380609035492,
      "learning_rate": 3.8099842737211734e-07,
      "loss": 0.3768,
      "num_input_tokens_seen": 23695544,
      "step": 36250
    },
    {
      "epoch": 19.0,
      "eval_loss": 0.47521016001701355,
      "eval_runtime": 9.7498,
      "eval_samples_per_second": 86.977,
      "eval_steps_per_second": 21.744,
      "num_input_tokens_seen": 23696296,
      "step": 36252
    },
    {
      "epoch": 19.001572327044027,
      "grad_norm": 0.409880667924881,
      "learning_rate": 3.7901236962379627e-07,
      "loss": 0.4987,
      "num_input_tokens_seen": 23697768,
      "step": 36255
    },
    {
      "epoch": 19.0041928721174,
      "grad_norm": 0.5655927062034607,
      "learning_rate": 3.7703146230519126e-07,
      "loss": 0.4218,
      "num_input_tokens_seen": 23700072,
      "step": 36260
    },
    {
      "epoch": 19.006813417190777,
      "grad_norm": 0.4144664406776428,
      "learning_rate": 3.7505570583069306e-07,
      "loss": 0.5444,
      "num_input_tokens_seen": 23702856,
      "step": 36265
    },
    {
      "epoch": 19.00943396226415,
      "grad_norm": 0.706672191619873,
      "learning_rate": 3.7308510061360435e-07,
      "loss": 0.4783,
      "num_input_tokens_seen": 23706184,
      "step": 36270
    },
    {
      "epoch": 19.012054507337528,
      "grad_norm": 0.49217596650123596,
      "learning_rate": 3.7111964706615374e-07,
      "loss": 0.466,
      "num_input_tokens_seen": 23709128,
      "step": 36275
    },
    {
      "epoch": 19.0146750524109,
      "grad_norm": 0.8090829849243164,
      "learning_rate": 3.691593455994929e-07,
      "loss": 0.6097,
      "num_input_tokens_seen": 23712808,
      "step": 36280
    },
    {
      "epoch": 19.017295597484278,
      "grad_norm": 0.3471272587776184,
      "learning_rate": 3.672041966236911e-07,
      "loss": 0.4153,
      "num_input_tokens_seen": 23716168,
      "step": 36285
    },
    {
      "epoch": 19.01991614255765,
      "grad_norm": 0.3089413046836853,
      "learning_rate": 3.6525420054774884e-07,
      "loss": 0.3562,
      "num_input_tokens_seen": 23719688,
      "step": 36290
    },
    {
      "epoch": 19.02253668763103,
      "grad_norm": 0.30022740364074707,
      "learning_rate": 3.6330935777957597e-07,
      "loss": 0.408,
      "num_input_tokens_seen": 23723752,
      "step": 36295
    },
    {
      "epoch": 19.0251572327044,
      "grad_norm": 0.40935108065605164,
      "learning_rate": 3.613696687260193e-07,
      "loss": 0.3599,
      "num_input_tokens_seen": 23727432,
      "step": 36300
    },
    {
      "epoch": 19.02777777777778,
      "grad_norm": 0.5868943929672241,
      "learning_rate": 3.59435133792832e-07,
      "loss": 0.4295,
      "num_input_tokens_seen": 23730312,
      "step": 36305
    },
    {
      "epoch": 19.030398322851152,
      "grad_norm": 0.5507081151008606,
      "learning_rate": 3.575057533847043e-07,
      "loss": 0.4831,
      "num_input_tokens_seen": 23733448,
      "step": 36310
    },
    {
      "epoch": 19.03301886792453,
      "grad_norm": 0.47393062710762024,
      "learning_rate": 3.555815279052327e-07,
      "loss": 0.4458,
      "num_input_tokens_seen": 23737320,
      "step": 36315
    },
    {
      "epoch": 19.035639412997902,
      "grad_norm": 0.3359716236591339,
      "learning_rate": 3.5366245775694805e-07,
      "loss": 0.5092,
      "num_input_tokens_seen": 23739752,
      "step": 36320
    },
    {
      "epoch": 19.03825995807128,
      "grad_norm": 0.6828017830848694,
      "learning_rate": 3.517485433412987e-07,
      "loss": 0.3528,
      "num_input_tokens_seen": 23742152,
      "step": 36325
    },
    {
      "epoch": 19.040880503144653,
      "grad_norm": 0.49744144082069397,
      "learning_rate": 3.4983978505865044e-07,
      "loss": 0.3944,
      "num_input_tokens_seen": 23745448,
      "step": 36330
    },
    {
      "epoch": 19.04350104821803,
      "grad_norm": 0.5325940251350403,
      "learning_rate": 3.47936183308295e-07,
      "loss": 0.4012,
      "num_input_tokens_seen": 23749864,
      "step": 36335
    },
    {
      "epoch": 19.046121593291403,
      "grad_norm": 0.49783799052238464,
      "learning_rate": 3.460377384884472e-07,
      "loss": 0.4658,
      "num_input_tokens_seen": 23752552,
      "step": 36340
    },
    {
      "epoch": 19.04874213836478,
      "grad_norm": 0.7018680572509766,
      "learning_rate": 3.4414445099623384e-07,
      "loss": 0.3638,
      "num_input_tokens_seen": 23755528,
      "step": 36345
    },
    {
      "epoch": 19.051362683438157,
      "grad_norm": 0.45031026005744934,
      "learning_rate": 3.42256321227713e-07,
      "loss": 0.3294,
      "num_input_tokens_seen": 23758600,
      "step": 36350
    },
    {
      "epoch": 19.05398322851153,
      "grad_norm": 0.8054553270339966,
      "learning_rate": 3.403733495778605e-07,
      "loss": 0.4351,
      "num_input_tokens_seen": 23761480,
      "step": 36355
    },
    {
      "epoch": 19.056603773584907,
      "grad_norm": 0.591643750667572,
      "learning_rate": 3.3849553644056955e-07,
      "loss": 0.4039,
      "num_input_tokens_seen": 23764168,
      "step": 36360
    },
    {
      "epoch": 19.05922431865828,
      "grad_norm": 0.8194330334663391,
      "learning_rate": 3.36622882208662e-07,
      "loss": 0.4971,
      "num_input_tokens_seen": 23766920,
      "step": 36365
    },
    {
      "epoch": 19.061844863731658,
      "grad_norm": 0.8816757798194885,
      "learning_rate": 3.347553872738718e-07,
      "loss": 0.4481,
      "num_input_tokens_seen": 23770056,
      "step": 36370
    },
    {
      "epoch": 19.06446540880503,
      "grad_norm": 0.43568000197410583,
      "learning_rate": 3.3289305202685583e-07,
      "loss": 0.316,
      "num_input_tokens_seen": 23773096,
      "step": 36375
    },
    {
      "epoch": 19.067085953878408,
      "grad_norm": 0.9880343675613403,
      "learning_rate": 3.310358768571997e-07,
      "loss": 0.3882,
      "num_input_tokens_seen": 23775496,
      "step": 36380
    },
    {
      "epoch": 19.06970649895178,
      "grad_norm": 0.5873074531555176,
      "learning_rate": 3.2918386215339827e-07,
      "loss": 0.4545,
      "num_input_tokens_seen": 23778600,
      "step": 36385
    },
    {
      "epoch": 19.072327044025158,
      "grad_norm": 0.1393464356660843,
      "learning_rate": 3.2733700830287205e-07,
      "loss": 0.4067,
      "num_input_tokens_seen": 23783688,
      "step": 36390
    },
    {
      "epoch": 19.07494758909853,
      "grad_norm": 0.5101221203804016,
      "learning_rate": 3.254953156919649e-07,
      "loss": 0.5105,
      "num_input_tokens_seen": 23786440,
      "step": 36395
    },
    {
      "epoch": 19.07756813417191,
      "grad_norm": 0.546928346157074,
      "learning_rate": 3.2365878470593527e-07,
      "loss": 0.5049,
      "num_input_tokens_seen": 23789768,
      "step": 36400
    },
    {
      "epoch": 19.080188679245282,
      "grad_norm": 0.5371906161308289,
      "learning_rate": 3.2182741572896747e-07,
      "loss": 0.4687,
      "num_input_tokens_seen": 23792264,
      "step": 36405
    },
    {
      "epoch": 19.08280922431866,
      "grad_norm": 0.5751212239265442,
      "learning_rate": 3.2000120914415786e-07,
      "loss": 0.4153,
      "num_input_tokens_seen": 23795816,
      "step": 36410
    },
    {
      "epoch": 19.085429769392032,
      "grad_norm": 0.43026235699653625,
      "learning_rate": 3.181801653335315e-07,
      "loss": 0.4933,
      "num_input_tokens_seen": 23798984,
      "step": 36415
    },
    {
      "epoch": 19.08805031446541,
      "grad_norm": 0.35377681255340576,
      "learning_rate": 3.163642846780307e-07,
      "loss": 0.424,
      "num_input_tokens_seen": 23802056,
      "step": 36420
    },
    {
      "epoch": 19.090670859538783,
      "grad_norm": 0.6219659447669983,
      "learning_rate": 3.1455356755751565e-07,
      "loss": 0.3971,
      "num_input_tokens_seen": 23804424,
      "step": 36425
    },
    {
      "epoch": 19.09329140461216,
      "grad_norm": 0.40414959192276,
      "learning_rate": 3.127480143507666e-07,
      "loss": 0.3633,
      "num_input_tokens_seen": 23807112,
      "step": 36430
    },
    {
      "epoch": 19.095911949685533,
      "grad_norm": 0.5278609991073608,
      "learning_rate": 3.1094762543548704e-07,
      "loss": 0.3897,
      "num_input_tokens_seen": 23810184,
      "step": 36435
    },
    {
      "epoch": 19.09853249475891,
      "grad_norm": 0.5099856853485107,
      "learning_rate": 3.091524011882951e-07,
      "loss": 0.4078,
      "num_input_tokens_seen": 23812776,
      "step": 36440
    },
    {
      "epoch": 19.101153039832287,
      "grad_norm": 0.4835525155067444,
      "learning_rate": 3.073623419847349e-07,
      "loss": 0.5297,
      "num_input_tokens_seen": 23816296,
      "step": 36445
    },
    {
      "epoch": 19.10377358490566,
      "grad_norm": 0.4336748421192169,
      "learning_rate": 3.0557744819926235e-07,
      "loss": 0.4854,
      "num_input_tokens_seen": 23819432,
      "step": 36450
    },
    {
      "epoch": 19.106394129979037,
      "grad_norm": 0.7134718298912048,
      "learning_rate": 3.037977202052594e-07,
      "loss": 0.5117,
      "num_input_tokens_seen": 23822600,
      "step": 36455
    },
    {
      "epoch": 19.10901467505241,
      "grad_norm": 0.4925031363964081,
      "learning_rate": 3.020231583750255e-07,
      "loss": 0.4163,
      "num_input_tokens_seen": 23825512,
      "step": 36460
    },
    {
      "epoch": 19.111635220125788,
      "grad_norm": 0.36237505078315735,
      "learning_rate": 3.002537630797747e-07,
      "loss": 0.4493,
      "num_input_tokens_seen": 23828456,
      "step": 36465
    },
    {
      "epoch": 19.11425576519916,
      "grad_norm": 0.3858928680419922,
      "learning_rate": 2.984895346896499e-07,
      "loss": 0.3633,
      "num_input_tokens_seen": 23831656,
      "step": 36470
    },
    {
      "epoch": 19.116876310272538,
      "grad_norm": 0.7376611828804016,
      "learning_rate": 2.967304735737031e-07,
      "loss": 0.4055,
      "num_input_tokens_seen": 23834312,
      "step": 36475
    },
    {
      "epoch": 19.11949685534591,
      "grad_norm": 0.5081597566604614,
      "learning_rate": 2.9497658009991214e-07,
      "loss": 0.454,
      "num_input_tokens_seen": 23836968,
      "step": 36480
    },
    {
      "epoch": 19.122117400419288,
      "grad_norm": 0.5356678366661072,
      "learning_rate": 2.932278546351724e-07,
      "loss": 0.4318,
      "num_input_tokens_seen": 23840616,
      "step": 36485
    },
    {
      "epoch": 19.12473794549266,
      "grad_norm": 0.6564772129058838,
      "learning_rate": 2.9148429754529137e-07,
      "loss": 0.411,
      "num_input_tokens_seen": 23844168,
      "step": 36490
    },
    {
      "epoch": 19.12735849056604,
      "grad_norm": 0.5905703902244568,
      "learning_rate": 2.897459091950078e-07,
      "loss": 0.3713,
      "num_input_tokens_seen": 23846888,
      "step": 36495
    },
    {
      "epoch": 19.129979035639412,
      "grad_norm": 0.3022751212120056,
      "learning_rate": 2.880126899479724e-07,
      "loss": 0.4945,
      "num_input_tokens_seen": 23850856,
      "step": 36500
    },
    {
      "epoch": 19.13259958071279,
      "grad_norm": 0.45928308367729187,
      "learning_rate": 2.862846401667507e-07,
      "loss": 0.4338,
      "num_input_tokens_seen": 23853992,
      "step": 36505
    },
    {
      "epoch": 19.135220125786162,
      "grad_norm": 0.5791561603546143,
      "learning_rate": 2.8456176021283697e-07,
      "loss": 0.4346,
      "num_input_tokens_seen": 23857864,
      "step": 36510
    },
    {
      "epoch": 19.13784067085954,
      "grad_norm": 0.47979921102523804,
      "learning_rate": 2.8284405044662897e-07,
      "loss": 0.4505,
      "num_input_tokens_seen": 23861480,
      "step": 36515
    },
    {
      "epoch": 19.140461215932913,
      "grad_norm": 0.692267894744873,
      "learning_rate": 2.8113151122745874e-07,
      "loss": 0.418,
      "num_input_tokens_seen": 23864424,
      "step": 36520
    },
    {
      "epoch": 19.14308176100629,
      "grad_norm": 0.6394004225730896,
      "learning_rate": 2.7942414291357023e-07,
      "loss": 0.5305,
      "num_input_tokens_seen": 23867272,
      "step": 36525
    },
    {
      "epoch": 19.145702306079663,
      "grad_norm": 0.5508009195327759,
      "learning_rate": 2.777219458621222e-07,
      "loss": 0.3761,
      "num_input_tokens_seen": 23871272,
      "step": 36530
    },
    {
      "epoch": 19.14832285115304,
      "grad_norm": 0.8453083634376526,
      "learning_rate": 2.7602492042919923e-07,
      "loss": 0.4119,
      "num_input_tokens_seen": 23874376,
      "step": 36535
    },
    {
      "epoch": 19.150943396226417,
      "grad_norm": 0.3748527765274048,
      "learning_rate": 2.7433306696979523e-07,
      "loss": 0.3657,
      "num_input_tokens_seen": 23877096,
      "step": 36540
    },
    {
      "epoch": 19.15356394129979,
      "grad_norm": 0.7263193130493164,
      "learning_rate": 2.7264638583782973e-07,
      "loss": 0.3646,
      "num_input_tokens_seen": 23880776,
      "step": 36545
    },
    {
      "epoch": 19.156184486373167,
      "grad_norm": 0.3670332431793213,
      "learning_rate": 2.7096487738613177e-07,
      "loss": 0.3024,
      "num_input_tokens_seen": 23885000,
      "step": 36550
    },
    {
      "epoch": 19.15880503144654,
      "grad_norm": 0.5246900916099548,
      "learning_rate": 2.6928854196646147e-07,
      "loss": 0.3554,
      "num_input_tokens_seen": 23887720,
      "step": 36555
    },
    {
      "epoch": 19.161425576519918,
      "grad_norm": 0.5146167278289795,
      "learning_rate": 2.6761737992948565e-07,
      "loss": 0.3936,
      "num_input_tokens_seen": 23890760,
      "step": 36560
    },
    {
      "epoch": 19.16404612159329,
      "grad_norm": 0.4351414144039154,
      "learning_rate": 2.6595139162479134e-07,
      "loss": 0.3822,
      "num_input_tokens_seen": 23893640,
      "step": 36565
    },
    {
      "epoch": 19.166666666666668,
      "grad_norm": 0.3233039677143097,
      "learning_rate": 2.6429057740088303e-07,
      "loss": 0.3758,
      "num_input_tokens_seen": 23896584,
      "step": 36570
    },
    {
      "epoch": 19.16928721174004,
      "grad_norm": 0.29783138632774353,
      "learning_rate": 2.6263493760518564e-07,
      "loss": 0.4969,
      "num_input_tokens_seen": 23901480,
      "step": 36575
    },
    {
      "epoch": 19.171907756813418,
      "grad_norm": 0.38340064883232117,
      "learning_rate": 2.6098447258404157e-07,
      "loss": 0.4649,
      "num_input_tokens_seen": 23904680,
      "step": 36580
    },
    {
      "epoch": 19.17452830188679,
      "grad_norm": 0.8125801086425781,
      "learning_rate": 2.593391826827107e-07,
      "loss": 0.3038,
      "num_input_tokens_seen": 23907144,
      "step": 36585
    },
    {
      "epoch": 19.17714884696017,
      "grad_norm": 0.4968878924846649,
      "learning_rate": 2.5769906824536225e-07,
      "loss": 0.4256,
      "num_input_tokens_seen": 23910888,
      "step": 36590
    },
    {
      "epoch": 19.179769392033542,
      "grad_norm": 0.6512069702148438,
      "learning_rate": 2.560641296150967e-07,
      "loss": 0.5746,
      "num_input_tokens_seen": 23914184,
      "step": 36595
    },
    {
      "epoch": 19.18238993710692,
      "grad_norm": 0.4289422631263733,
      "learning_rate": 2.544343671339211e-07,
      "loss": 0.4059,
      "num_input_tokens_seen": 23917608,
      "step": 36600
    },
    {
      "epoch": 19.185010482180292,
      "grad_norm": 0.5892848968505859,
      "learning_rate": 2.5280978114276264e-07,
      "loss": 0.4326,
      "num_input_tokens_seen": 23922312,
      "step": 36605
    },
    {
      "epoch": 19.18763102725367,
      "grad_norm": 0.6073488593101501,
      "learning_rate": 2.511903719814662e-07,
      "loss": 0.4996,
      "num_input_tokens_seen": 23925096,
      "step": 36610
    },
    {
      "epoch": 19.190251572327043,
      "grad_norm": 0.7643356323242188,
      "learning_rate": 2.4957613998879695e-07,
      "loss": 0.4514,
      "num_input_tokens_seen": 23927976,
      "step": 36615
    },
    {
      "epoch": 19.19287211740042,
      "grad_norm": 0.4403240382671356,
      "learning_rate": 2.479670855024291e-07,
      "loss": 0.4032,
      "num_input_tokens_seen": 23934760,
      "step": 36620
    },
    {
      "epoch": 19.195492662473793,
      "grad_norm": 0.4208593964576721,
      "learning_rate": 2.463632088589629e-07,
      "loss": 0.4669,
      "num_input_tokens_seen": 23937960,
      "step": 36625
    },
    {
      "epoch": 19.19811320754717,
      "grad_norm": 0.39962759613990784,
      "learning_rate": 2.447645103939078e-07,
      "loss": 0.3588,
      "num_input_tokens_seen": 23940904,
      "step": 36630
    },
    {
      "epoch": 19.200733752620547,
      "grad_norm": 0.5633230805397034,
      "learning_rate": 2.431709904416962e-07,
      "loss": 0.385,
      "num_input_tokens_seen": 23943848,
      "step": 36635
    },
    {
      "epoch": 19.20335429769392,
      "grad_norm": 0.5279536247253418,
      "learning_rate": 2.4158264933567264e-07,
      "loss": 0.4242,
      "num_input_tokens_seen": 23947112,
      "step": 36640
    },
    {
      "epoch": 19.205974842767297,
      "grad_norm": 0.6001156568527222,
      "learning_rate": 2.399994874080963e-07,
      "loss": 0.3444,
      "num_input_tokens_seen": 23950056,
      "step": 36645
    },
    {
      "epoch": 19.20859538784067,
      "grad_norm": 0.615225076675415,
      "learning_rate": 2.3842150499015504e-07,
      "loss": 0.4696,
      "num_input_tokens_seen": 23953320,
      "step": 36650
    },
    {
      "epoch": 19.211215932914047,
      "grad_norm": 0.5663082599639893,
      "learning_rate": 2.368487024119348e-07,
      "loss": 0.4437,
      "num_input_tokens_seen": 23956168,
      "step": 36655
    },
    {
      "epoch": 19.21383647798742,
      "grad_norm": 0.42328500747680664,
      "learning_rate": 2.3528108000245573e-07,
      "loss": 0.4088,
      "num_input_tokens_seen": 23959528,
      "step": 36660
    },
    {
      "epoch": 19.216457023060798,
      "grad_norm": 0.40425625443458557,
      "learning_rate": 2.3371863808964445e-07,
      "loss": 0.4435,
      "num_input_tokens_seen": 23962504,
      "step": 36665
    },
    {
      "epoch": 19.21907756813417,
      "grad_norm": 0.4231935441493988,
      "learning_rate": 2.32161377000345e-07,
      "loss": 0.3731,
      "num_input_tokens_seen": 23965736,
      "step": 36670
    },
    {
      "epoch": 19.221698113207548,
      "grad_norm": 0.46732401847839355,
      "learning_rate": 2.306092970603163e-07,
      "loss": 0.3596,
      "num_input_tokens_seen": 23968328,
      "step": 36675
    },
    {
      "epoch": 19.22431865828092,
      "grad_norm": 0.8950494527816772,
      "learning_rate": 2.2906239859424306e-07,
      "loss": 0.388,
      "num_input_tokens_seen": 23972360,
      "step": 36680
    },
    {
      "epoch": 19.2269392033543,
      "grad_norm": 0.5742785930633545,
      "learning_rate": 2.2752068192571084e-07,
      "loss": 0.4302,
      "num_input_tokens_seen": 23975688,
      "step": 36685
    },
    {
      "epoch": 19.229559748427672,
      "grad_norm": 0.5517892837524414,
      "learning_rate": 2.2598414737723396e-07,
      "loss": 0.4851,
      "num_input_tokens_seen": 23979656,
      "step": 36690
    },
    {
      "epoch": 19.23218029350105,
      "grad_norm": 0.6474600434303284,
      "learning_rate": 2.2445279527023578e-07,
      "loss": 0.4721,
      "num_input_tokens_seen": 23982568,
      "step": 36695
    },
    {
      "epoch": 19.234800838574422,
      "grad_norm": 0.39018872380256653,
      "learning_rate": 2.2292662592506007e-07,
      "loss": 0.4868,
      "num_input_tokens_seen": 23985640,
      "step": 36700
    },
    {
      "epoch": 19.2374213836478,
      "grad_norm": 0.45002856850624084,
      "learning_rate": 2.2140563966095985e-07,
      "loss": 0.4764,
      "num_input_tokens_seen": 23989000,
      "step": 36705
    },
    {
      "epoch": 19.240041928721173,
      "grad_norm": 0.3823266327381134,
      "learning_rate": 2.1988983679611386e-07,
      "loss": 0.4434,
      "num_input_tokens_seen": 23991784,
      "step": 36710
    },
    {
      "epoch": 19.24266247379455,
      "grad_norm": 0.613291323184967,
      "learning_rate": 2.1837921764760462e-07,
      "loss": 0.4087,
      "num_input_tokens_seen": 23996328,
      "step": 36715
    },
    {
      "epoch": 19.245283018867923,
      "grad_norm": 0.6284399032592773,
      "learning_rate": 2.1687378253144043e-07,
      "loss": 0.5159,
      "num_input_tokens_seen": 23999816,
      "step": 36720
    },
    {
      "epoch": 19.2479035639413,
      "grad_norm": 0.44724708795547485,
      "learning_rate": 2.153735317625416e-07,
      "loss": 0.4818,
      "num_input_tokens_seen": 24002472,
      "step": 36725
    },
    {
      "epoch": 19.250524109014677,
      "grad_norm": 0.37409457564353943,
      "learning_rate": 2.1387846565474045e-07,
      "loss": 0.2916,
      "num_input_tokens_seen": 24005256,
      "step": 36730
    },
    {
      "epoch": 19.25314465408805,
      "grad_norm": 0.3896079659461975,
      "learning_rate": 2.123885845207896e-07,
      "loss": 0.4169,
      "num_input_tokens_seen": 24008776,
      "step": 36735
    },
    {
      "epoch": 19.255765199161427,
      "grad_norm": 0.445680171251297,
      "learning_rate": 2.1090388867235356e-07,
      "loss": 0.3313,
      "num_input_tokens_seen": 24012904,
      "step": 36740
    },
    {
      "epoch": 19.2583857442348,
      "grad_norm": 0.9187255501747131,
      "learning_rate": 2.0942437842001728e-07,
      "loss": 0.392,
      "num_input_tokens_seen": 24016040,
      "step": 36745
    },
    {
      "epoch": 19.261006289308177,
      "grad_norm": 0.42840757966041565,
      "learning_rate": 2.0795005407327483e-07,
      "loss": 0.4397,
      "num_input_tokens_seen": 24018760,
      "step": 36750
    },
    {
      "epoch": 19.26362683438155,
      "grad_norm": 0.2893737554550171,
      "learning_rate": 2.0648091594053508e-07,
      "loss": 0.4753,
      "num_input_tokens_seen": 24022088,
      "step": 36755
    },
    {
      "epoch": 19.266247379454928,
      "grad_norm": 0.7463483810424805,
      "learning_rate": 2.0501696432913275e-07,
      "loss": 0.4067,
      "num_input_tokens_seen": 24025064,
      "step": 36760
    },
    {
      "epoch": 19.2688679245283,
      "grad_norm": 0.40210670232772827,
      "learning_rate": 2.035581995453034e-07,
      "loss": 0.4162,
      "num_input_tokens_seen": 24028264,
      "step": 36765
    },
    {
      "epoch": 19.271488469601678,
      "grad_norm": 0.31510820984840393,
      "learning_rate": 2.0210462189420577e-07,
      "loss": 0.4724,
      "num_input_tokens_seen": 24031560,
      "step": 36770
    },
    {
      "epoch": 19.27410901467505,
      "grad_norm": 0.2667003870010376,
      "learning_rate": 2.006562316799132e-07,
      "loss": 0.4924,
      "num_input_tokens_seen": 24035944,
      "step": 36775
    },
    {
      "epoch": 19.27672955974843,
      "grad_norm": 0.5444145202636719,
      "learning_rate": 1.992130292054084e-07,
      "loss": 0.6033,
      "num_input_tokens_seen": 24039912,
      "step": 36780
    },
    {
      "epoch": 19.279350104821802,
      "grad_norm": 0.5201069712638855,
      "learning_rate": 1.9777501477259708e-07,
      "loss": 0.3655,
      "num_input_tokens_seen": 24042792,
      "step": 36785
    },
    {
      "epoch": 19.28197064989518,
      "grad_norm": 0.500718355178833,
      "learning_rate": 1.9634218868229414e-07,
      "loss": 0.4335,
      "num_input_tokens_seen": 24046184,
      "step": 36790
    },
    {
      "epoch": 19.284591194968552,
      "grad_norm": 0.4904308617115021,
      "learning_rate": 1.9491455123423207e-07,
      "loss": 0.4555,
      "num_input_tokens_seen": 24049576,
      "step": 36795
    },
    {
      "epoch": 19.28721174004193,
      "grad_norm": 0.5565455555915833,
      "learning_rate": 1.9349210272705253e-07,
      "loss": 0.3928,
      "num_input_tokens_seen": 24051944,
      "step": 36800
    },
    {
      "epoch": 19.289832285115303,
      "grad_norm": 0.3078770041465759,
      "learning_rate": 1.920748434583175e-07,
      "loss": 0.4784,
      "num_input_tokens_seen": 24055112,
      "step": 36805
    },
    {
      "epoch": 19.29245283018868,
      "grad_norm": 0.5365417003631592,
      "learning_rate": 1.906627737245009e-07,
      "loss": 0.4418,
      "num_input_tokens_seen": 24058888,
      "step": 36810
    },
    {
      "epoch": 19.295073375262053,
      "grad_norm": 0.25484785437583923,
      "learning_rate": 1.8925589382099152e-07,
      "loss": 0.3904,
      "num_input_tokens_seen": 24062728,
      "step": 36815
    },
    {
      "epoch": 19.29769392033543,
      "grad_norm": 0.2793627381324768,
      "learning_rate": 1.8785420404209553e-07,
      "loss": 0.6113,
      "num_input_tokens_seen": 24066504,
      "step": 36820
    },
    {
      "epoch": 19.300314465408803,
      "grad_norm": 0.45662426948547363,
      "learning_rate": 1.864577046810284e-07,
      "loss": 0.3991,
      "num_input_tokens_seen": 24069256,
      "step": 36825
    },
    {
      "epoch": 19.30293501048218,
      "grad_norm": 0.4276115298271179,
      "learning_rate": 1.8506639602992038e-07,
      "loss": 0.3964,
      "num_input_tokens_seen": 24071720,
      "step": 36830
    },
    {
      "epoch": 19.305555555555557,
      "grad_norm": 0.28666582703590393,
      "learning_rate": 1.836802783798164e-07,
      "loss": 0.2941,
      "num_input_tokens_seen": 24075720,
      "step": 36835
    },
    {
      "epoch": 19.30817610062893,
      "grad_norm": 0.5867999792098999,
      "learning_rate": 1.822993520206817e-07,
      "loss": 0.367,
      "num_input_tokens_seen": 24078920,
      "step": 36840
    },
    {
      "epoch": 19.310796645702307,
      "grad_norm": 0.49456408619880676,
      "learning_rate": 1.8092361724138528e-07,
      "loss": 0.4179,
      "num_input_tokens_seen": 24082376,
      "step": 36845
    },
    {
      "epoch": 19.31341719077568,
      "grad_norm": 0.26395246386528015,
      "learning_rate": 1.7955307432971914e-07,
      "loss": 0.3792,
      "num_input_tokens_seen": 24085928,
      "step": 36850
    },
    {
      "epoch": 19.316037735849058,
      "grad_norm": 0.6177346706390381,
      "learning_rate": 1.7818772357238445e-07,
      "loss": 0.4272,
      "num_input_tokens_seen": 24088648,
      "step": 36855
    },
    {
      "epoch": 19.31865828092243,
      "grad_norm": 0.5218931436538696,
      "learning_rate": 1.7682756525499166e-07,
      "loss": 0.3622,
      "num_input_tokens_seen": 24092136,
      "step": 36860
    },
    {
      "epoch": 19.321278825995808,
      "grad_norm": 0.6452751159667969,
      "learning_rate": 1.7547259966207708e-07,
      "loss": 0.4952,
      "num_input_tokens_seen": 24094632,
      "step": 36865
    },
    {
      "epoch": 19.32389937106918,
      "grad_norm": 0.49546492099761963,
      "learning_rate": 1.7412282707708062e-07,
      "loss": 0.4643,
      "num_input_tokens_seen": 24098248,
      "step": 36870
    },
    {
      "epoch": 19.32651991614256,
      "grad_norm": 0.5519476532936096,
      "learning_rate": 1.7277824778235974e-07,
      "loss": 0.4447,
      "num_input_tokens_seen": 24100872,
      "step": 36875
    },
    {
      "epoch": 19.329140461215932,
      "grad_norm": 0.563471794128418,
      "learning_rate": 1.714388620591867e-07,
      "loss": 0.5099,
      "num_input_tokens_seen": 24104520,
      "step": 36880
    },
    {
      "epoch": 19.33176100628931,
      "grad_norm": 0.0893794521689415,
      "learning_rate": 1.701046701877429e-07,
      "loss": 0.4417,
      "num_input_tokens_seen": 24111048,
      "step": 36885
    },
    {
      "epoch": 19.334381551362682,
      "grad_norm": 0.526986837387085,
      "learning_rate": 1.6877567244713011e-07,
      "loss": 0.4282,
      "num_input_tokens_seen": 24113448,
      "step": 36890
    },
    {
      "epoch": 19.33700209643606,
      "grad_norm": 0.6591149568557739,
      "learning_rate": 1.6745186911535372e-07,
      "loss": 0.4024,
      "num_input_tokens_seen": 24116264,
      "step": 36895
    },
    {
      "epoch": 19.339622641509433,
      "grad_norm": 0.5180711150169373,
      "learning_rate": 1.6613326046934497e-07,
      "loss": 0.4014,
      "num_input_tokens_seen": 24118888,
      "step": 36900
    },
    {
      "epoch": 19.34224318658281,
      "grad_norm": 0.2987484931945801,
      "learning_rate": 1.64819846784936e-07,
      "loss": 0.4775,
      "num_input_tokens_seen": 24122056,
      "step": 36905
    },
    {
      "epoch": 19.344863731656183,
      "grad_norm": 0.49039319157600403,
      "learning_rate": 1.6351162833688204e-07,
      "loss": 0.5093,
      "num_input_tokens_seen": 24125288,
      "step": 36910
    },
    {
      "epoch": 19.34748427672956,
      "grad_norm": 0.4222317636013031,
      "learning_rate": 1.6220860539884475e-07,
      "loss": 0.5065,
      "num_input_tokens_seen": 24128904,
      "step": 36915
    },
    {
      "epoch": 19.350104821802937,
      "grad_norm": 0.4463026523590088,
      "learning_rate": 1.6091077824340327e-07,
      "loss": 0.5227,
      "num_input_tokens_seen": 24131880,
      "step": 36920
    },
    {
      "epoch": 19.35272536687631,
      "grad_norm": 0.6320311427116394,
      "learning_rate": 1.5961814714204605e-07,
      "loss": 0.414,
      "num_input_tokens_seen": 24134600,
      "step": 36925
    },
    {
      "epoch": 19.355345911949687,
      "grad_norm": 0.25533953309059143,
      "learning_rate": 1.5833071236518172e-07,
      "loss": 0.3986,
      "num_input_tokens_seen": 24137448,
      "step": 36930
    },
    {
      "epoch": 19.35796645702306,
      "grad_norm": 0.38764870166778564,
      "learning_rate": 1.5704847418212266e-07,
      "loss": 0.5902,
      "num_input_tokens_seen": 24141064,
      "step": 36935
    },
    {
      "epoch": 19.360587002096437,
      "grad_norm": 0.3795212507247925,
      "learning_rate": 1.5577143286109875e-07,
      "loss": 0.3595,
      "num_input_tokens_seen": 24143688,
      "step": 36940
    },
    {
      "epoch": 19.36320754716981,
      "grad_norm": 0.5601648688316345,
      "learning_rate": 1.5449958866925185e-07,
      "loss": 0.304,
      "num_input_tokens_seen": 24146984,
      "step": 36945
    },
    {
      "epoch": 19.365828092243188,
      "grad_norm": 0.39265164732933044,
      "learning_rate": 1.532329418726386e-07,
      "loss": 0.5304,
      "num_input_tokens_seen": 24151784,
      "step": 36950
    },
    {
      "epoch": 19.36844863731656,
      "grad_norm": 0.33940061926841736,
      "learning_rate": 1.5197149273623036e-07,
      "loss": 0.4546,
      "num_input_tokens_seen": 24154792,
      "step": 36955
    },
    {
      "epoch": 19.371069182389938,
      "grad_norm": 1.305700659751892,
      "learning_rate": 1.5071524152389936e-07,
      "loss": 0.468,
      "num_input_tokens_seen": 24157512,
      "step": 36960
    },
    {
      "epoch": 19.37368972746331,
      "grad_norm": 0.38468384742736816,
      "learning_rate": 1.4946418849844934e-07,
      "loss": 0.5373,
      "num_input_tokens_seen": 24160392,
      "step": 36965
    },
    {
      "epoch": 19.37631027253669,
      "grad_norm": 0.7614043354988098,
      "learning_rate": 1.4821833392157923e-07,
      "loss": 0.4106,
      "num_input_tokens_seen": 24163656,
      "step": 36970
    },
    {
      "epoch": 19.378930817610062,
      "grad_norm": 0.3918688893318176,
      "learning_rate": 1.4697767805390838e-07,
      "loss": 0.4531,
      "num_input_tokens_seen": 24166632,
      "step": 36975
    },
    {
      "epoch": 19.38155136268344,
      "grad_norm": 0.6783296465873718,
      "learning_rate": 1.4574222115497084e-07,
      "loss": 0.4414,
      "num_input_tokens_seen": 24169320,
      "step": 36980
    },
    {
      "epoch": 19.384171907756812,
      "grad_norm": 0.5758150219917297,
      "learning_rate": 1.4451196348320706e-07,
      "loss": 0.5452,
      "num_input_tokens_seen": 24172520,
      "step": 36985
    },
    {
      "epoch": 19.38679245283019,
      "grad_norm": 0.5775917768478394,
      "learning_rate": 1.4328690529597232e-07,
      "loss": 0.333,
      "num_input_tokens_seen": 24176232,
      "step": 36990
    },
    {
      "epoch": 19.389412997903563,
      "grad_norm": 0.4894263446331024,
      "learning_rate": 1.4206704684953943e-07,
      "loss": 0.448,
      "num_input_tokens_seen": 24179400,
      "step": 36995
    },
    {
      "epoch": 19.39203354297694,
      "grad_norm": 0.8257578611373901,
      "learning_rate": 1.4085238839908476e-07,
      "loss": 0.4791,
      "num_input_tokens_seen": 24182568,
      "step": 37000
    },
    {
      "epoch": 19.394654088050313,
      "grad_norm": 0.542210042476654,
      "learning_rate": 1.3964293019870235e-07,
      "loss": 0.4492,
      "num_input_tokens_seen": 24185640,
      "step": 37005
    },
    {
      "epoch": 19.39727463312369,
      "grad_norm": 0.604803740978241,
      "learning_rate": 1.3843867250139807e-07,
      "loss": 0.494,
      "num_input_tokens_seen": 24189256,
      "step": 37010
    },
    {
      "epoch": 19.399895178197063,
      "grad_norm": 0.2723741829395294,
      "learning_rate": 1.3723961555908715e-07,
      "loss": 0.4076,
      "num_input_tokens_seen": 24192744,
      "step": 37015
    },
    {
      "epoch": 19.40251572327044,
      "grad_norm": 0.3658246695995331,
      "learning_rate": 1.3604575962259947e-07,
      "loss": 0.4929,
      "num_input_tokens_seen": 24197128,
      "step": 37020
    },
    {
      "epoch": 19.405136268343817,
      "grad_norm": 0.3507083058357239,
      "learning_rate": 1.3485710494167692e-07,
      "loss": 0.3619,
      "num_input_tokens_seen": 24200552,
      "step": 37025
    },
    {
      "epoch": 19.40775681341719,
      "grad_norm": 0.461526095867157,
      "learning_rate": 1.3367365176496783e-07,
      "loss": 0.4185,
      "num_input_tokens_seen": 24203880,
      "step": 37030
    },
    {
      "epoch": 19.410377358490567,
      "grad_norm": 0.5305723547935486,
      "learning_rate": 1.324954003400436e-07,
      "loss": 0.4738,
      "num_input_tokens_seen": 24207752,
      "step": 37035
    },
    {
      "epoch": 19.41299790356394,
      "grad_norm": 0.5751562714576721,
      "learning_rate": 1.313223509133793e-07,
      "loss": 0.5024,
      "num_input_tokens_seen": 24211752,
      "step": 37040
    },
    {
      "epoch": 19.415618448637318,
      "grad_norm": 0.6543489098548889,
      "learning_rate": 1.301545037303592e-07,
      "loss": 0.4029,
      "num_input_tokens_seen": 24213992,
      "step": 37045
    },
    {
      "epoch": 19.41823899371069,
      "grad_norm": 0.5826218724250793,
      "learning_rate": 1.2899185903529066e-07,
      "loss": 0.4904,
      "num_input_tokens_seen": 24216584,
      "step": 37050
    },
    {
      "epoch": 19.420859538784068,
      "grad_norm": 0.7983960509300232,
      "learning_rate": 1.278344170713791e-07,
      "loss": 0.4096,
      "num_input_tokens_seen": 24219336,
      "step": 37055
    },
    {
      "epoch": 19.42348008385744,
      "grad_norm": 0.972006618976593,
      "learning_rate": 1.2668217808075033e-07,
      "loss": 0.4119,
      "num_input_tokens_seen": 24222024,
      "step": 37060
    },
    {
      "epoch": 19.42610062893082,
      "grad_norm": 0.5430517196655273,
      "learning_rate": 1.255351423044393e-07,
      "loss": 0.4455,
      "num_input_tokens_seen": 24225384,
      "step": 37065
    },
    {
      "epoch": 19.428721174004192,
      "grad_norm": 0.5347617268562317,
      "learning_rate": 1.2439330998239572e-07,
      "loss": 0.4075,
      "num_input_tokens_seen": 24228232,
      "step": 37070
    },
    {
      "epoch": 19.43134171907757,
      "grad_norm": 0.6950091123580933,
      "learning_rate": 1.2325668135347578e-07,
      "loss": 0.498,
      "num_input_tokens_seen": 24231432,
      "step": 37075
    },
    {
      "epoch": 19.433962264150942,
      "grad_norm": 0.29892513155937195,
      "learning_rate": 1.221252566554476e-07,
      "loss": 0.3734,
      "num_input_tokens_seen": 24234952,
      "step": 37080
    },
    {
      "epoch": 19.43658280922432,
      "grad_norm": 0.5164527893066406,
      "learning_rate": 1.2099903612499685e-07,
      "loss": 0.3886,
      "num_input_tokens_seen": 24237768,
      "step": 37085
    },
    {
      "epoch": 19.439203354297693,
      "grad_norm": 0.29072707891464233,
      "learning_rate": 1.198780199977101e-07,
      "loss": 0.4956,
      "num_input_tokens_seen": 24241032,
      "step": 37090
    },
    {
      "epoch": 19.44182389937107,
      "grad_norm": 0.2876918315887451,
      "learning_rate": 1.1876220850809416e-07,
      "loss": 0.4756,
      "num_input_tokens_seen": 24245416,
      "step": 37095
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 0.9489084482192993,
      "learning_rate": 1.1765160188956514e-07,
      "loss": 0.4978,
      "num_input_tokens_seen": 24249128,
      "step": 37100
    },
    {
      "epoch": 19.44706498951782,
      "grad_norm": 0.3439037501811981,
      "learning_rate": 1.1654620037445107e-07,
      "loss": 0.3804,
      "num_input_tokens_seen": 24252200,
      "step": 37105
    },
    {
      "epoch": 19.449685534591197,
      "grad_norm": 0.3569785952568054,
      "learning_rate": 1.1544600419398088e-07,
      "loss": 0.4231,
      "num_input_tokens_seen": 24255112,
      "step": 37110
    },
    {
      "epoch": 19.45230607966457,
      "grad_norm": 0.45534855127334595,
      "learning_rate": 1.1435101357831213e-07,
      "loss": 0.5847,
      "num_input_tokens_seen": 24258504,
      "step": 37115
    },
    {
      "epoch": 19.454926624737947,
      "grad_norm": 0.6773272156715393,
      "learning_rate": 1.1326122875650047e-07,
      "loss": 0.475,
      "num_input_tokens_seen": 24260776,
      "step": 37120
    },
    {
      "epoch": 19.45754716981132,
      "grad_norm": 0.8609417676925659,
      "learning_rate": 1.1217664995651633e-07,
      "loss": 0.3722,
      "num_input_tokens_seen": 24263976,
      "step": 37125
    },
    {
      "epoch": 19.460167714884697,
      "grad_norm": 0.5769352316856384,
      "learning_rate": 1.1109727740523935e-07,
      "loss": 0.4245,
      "num_input_tokens_seen": 24267144,
      "step": 37130
    },
    {
      "epoch": 19.46278825995807,
      "grad_norm": 0.29699981212615967,
      "learning_rate": 1.1002311132846944e-07,
      "loss": 0.4354,
      "num_input_tokens_seen": 24270312,
      "step": 37135
    },
    {
      "epoch": 19.465408805031448,
      "grad_norm": 0.5206173062324524,
      "learning_rate": 1.0895415195090186e-07,
      "loss": 0.4082,
      "num_input_tokens_seen": 24274632,
      "step": 37140
    },
    {
      "epoch": 19.46802935010482,
      "grad_norm": 0.6135611534118652,
      "learning_rate": 1.0789039949615221e-07,
      "loss": 0.577,
      "num_input_tokens_seen": 24279144,
      "step": 37145
    },
    {
      "epoch": 19.470649895178198,
      "grad_norm": 0.8500182032585144,
      "learning_rate": 1.0683185418674801e-07,
      "loss": 0.484,
      "num_input_tokens_seen": 24282024,
      "step": 37150
    },
    {
      "epoch": 19.47327044025157,
      "grad_norm": 1.5980077981948853,
      "learning_rate": 1.0577851624412604e-07,
      "loss": 0.5095,
      "num_input_tokens_seen": 24285288,
      "step": 37155
    },
    {
      "epoch": 19.47589098532495,
      "grad_norm": 0.39529258012771606,
      "learning_rate": 1.0473038588862672e-07,
      "loss": 0.3604,
      "num_input_tokens_seen": 24289384,
      "step": 37160
    },
    {
      "epoch": 19.478511530398322,
      "grad_norm": 0.8391727805137634,
      "learning_rate": 1.0368746333951351e-07,
      "loss": 0.581,
      "num_input_tokens_seen": 24291848,
      "step": 37165
    },
    {
      "epoch": 19.4811320754717,
      "grad_norm": 0.5113101601600647,
      "learning_rate": 1.0264974881494805e-07,
      "loss": 0.3464,
      "num_input_tokens_seen": 24296296,
      "step": 37170
    },
    {
      "epoch": 19.483752620545072,
      "grad_norm": 0.42721453309059143,
      "learning_rate": 1.0161724253201222e-07,
      "loss": 0.6834,
      "num_input_tokens_seen": 24299176,
      "step": 37175
    },
    {
      "epoch": 19.48637316561845,
      "grad_norm": 0.4274456799030304,
      "learning_rate": 1.0058994470669436e-07,
      "loss": 0.3602,
      "num_input_tokens_seen": 24302216,
      "step": 37180
    },
    {
      "epoch": 19.488993710691823,
      "grad_norm": 0.5548375844955444,
      "learning_rate": 9.956785555388926e-08,
      "loss": 0.5161,
      "num_input_tokens_seen": 24304648,
      "step": 37185
    },
    {
      "epoch": 19.4916142557652,
      "grad_norm": 0.9742227792739868,
      "learning_rate": 9.855097528741475e-08,
      "loss": 0.4142,
      "num_input_tokens_seen": 24308712,
      "step": 37190
    },
    {
      "epoch": 19.494234800838573,
      "grad_norm": 0.5171747803688049,
      "learning_rate": 9.753930411998402e-08,
      "loss": 0.411,
      "num_input_tokens_seen": 24312072,
      "step": 37195
    },
    {
      "epoch": 19.49685534591195,
      "grad_norm": 0.2677444815635681,
      "learning_rate": 9.653284226322501e-08,
      "loss": 0.3572,
      "num_input_tokens_seen": 24318312,
      "step": 37200
    },
    {
      "epoch": 19.499475890985323,
      "grad_norm": 0.5015606880187988,
      "learning_rate": 9.553158992768318e-08,
      "loss": 0.4715,
      "num_input_tokens_seen": 24321128,
      "step": 37205
    },
    {
      "epoch": 19.5020964360587,
      "grad_norm": 0.42304298281669617,
      "learning_rate": 9.453554732281044e-08,
      "loss": 0.382,
      "num_input_tokens_seen": 24324392,
      "step": 37210
    },
    {
      "epoch": 19.504716981132077,
      "grad_norm": 3.078249216079712,
      "learning_rate": 9.354471465696236e-08,
      "loss": 0.6231,
      "num_input_tokens_seen": 24327560,
      "step": 37215
    },
    {
      "epoch": 19.50733752620545,
      "grad_norm": 0.3888397216796875,
      "learning_rate": 9.255909213741199e-08,
      "loss": 0.5343,
      "num_input_tokens_seen": 24330504,
      "step": 37220
    },
    {
      "epoch": 19.509958071278827,
      "grad_norm": 0.5247480273246765,
      "learning_rate": 9.157867997034164e-08,
      "loss": 0.3562,
      "num_input_tokens_seen": 24333800,
      "step": 37225
    },
    {
      "epoch": 19.5125786163522,
      "grad_norm": 0.49553442001342773,
      "learning_rate": 9.060347836084004e-08,
      "loss": 0.4231,
      "num_input_tokens_seen": 24336264,
      "step": 37230
    },
    {
      "epoch": 19.515199161425578,
      "grad_norm": 0.7437885999679565,
      "learning_rate": 8.96334875129079e-08,
      "loss": 0.5148,
      "num_input_tokens_seen": 24338696,
      "step": 37235
    },
    {
      "epoch": 19.51781970649895,
      "grad_norm": 0.7096288800239563,
      "learning_rate": 8.866870762946067e-08,
      "loss": 0.4085,
      "num_input_tokens_seen": 24342056,
      "step": 37240
    },
    {
      "epoch": 19.520440251572328,
      "grad_norm": 0.3684318959712982,
      "learning_rate": 8.77091389123147e-08,
      "loss": 0.4558,
      "num_input_tokens_seen": 24345480,
      "step": 37245
    },
    {
      "epoch": 19.5230607966457,
      "grad_norm": 0.36343175172805786,
      "learning_rate": 8.675478156220385e-08,
      "loss": 0.4462,
      "num_input_tokens_seen": 24348680,
      "step": 37250
    },
    {
      "epoch": 19.52568134171908,
      "grad_norm": 0.2937139868736267,
      "learning_rate": 8.580563577876566e-08,
      "loss": 0.3846,
      "num_input_tokens_seen": 24351624,
      "step": 37255
    },
    {
      "epoch": 19.528301886792452,
      "grad_norm": 0.35794511437416077,
      "learning_rate": 8.486170176055242e-08,
      "loss": 0.4651,
      "num_input_tokens_seen": 24356552,
      "step": 37260
    },
    {
      "epoch": 19.53092243186583,
      "grad_norm": 0.4716980755329132,
      "learning_rate": 8.392297970502282e-08,
      "loss": 0.3945,
      "num_input_tokens_seen": 24359688,
      "step": 37265
    },
    {
      "epoch": 19.533542976939202,
      "grad_norm": 0.34283822774887085,
      "learning_rate": 8.298946980855315e-08,
      "loss": 0.4343,
      "num_input_tokens_seen": 24364328,
      "step": 37270
    },
    {
      "epoch": 19.53616352201258,
      "grad_norm": 0.43918150663375854,
      "learning_rate": 8.206117226641497e-08,
      "loss": 0.3014,
      "num_input_tokens_seen": 24367752,
      "step": 37275
    },
    {
      "epoch": 19.538784067085953,
      "grad_norm": 0.4366559684276581,
      "learning_rate": 8.113808727280015e-08,
      "loss": 0.458,
      "num_input_tokens_seen": 24371208,
      "step": 37280
    },
    {
      "epoch": 19.54140461215933,
      "grad_norm": 0.5807639360427856,
      "learning_rate": 8.022021502080979e-08,
      "loss": 0.3503,
      "num_input_tokens_seen": 24374824,
      "step": 37285
    },
    {
      "epoch": 19.544025157232703,
      "grad_norm": 0.34020480513572693,
      "learning_rate": 7.930755570244863e-08,
      "loss": 0.3512,
      "num_input_tokens_seen": 24378408,
      "step": 37290
    },
    {
      "epoch": 19.54664570230608,
      "grad_norm": 0.7405974268913269,
      "learning_rate": 7.840010950863619e-08,
      "loss": 0.544,
      "num_input_tokens_seen": 24381160,
      "step": 37295
    },
    {
      "epoch": 19.549266247379457,
      "grad_norm": 0.6507527232170105,
      "learning_rate": 7.749787662920116e-08,
      "loss": 0.5857,
      "num_input_tokens_seen": 24384072,
      "step": 37300
    },
    {
      "epoch": 19.55188679245283,
      "grad_norm": 0.810911238193512,
      "learning_rate": 7.660085725288146e-08,
      "loss": 0.36,
      "num_input_tokens_seen": 24386952,
      "step": 37305
    },
    {
      "epoch": 19.554507337526207,
      "grad_norm": 0.34696516394615173,
      "learning_rate": 7.570905156731867e-08,
      "loss": 0.3809,
      "num_input_tokens_seen": 24390600,
      "step": 37310
    },
    {
      "epoch": 19.55712788259958,
      "grad_norm": 0.40827107429504395,
      "learning_rate": 7.482245975907188e-08,
      "loss": 0.4289,
      "num_input_tokens_seen": 24393416,
      "step": 37315
    },
    {
      "epoch": 19.559748427672957,
      "grad_norm": 0.6325499415397644,
      "learning_rate": 7.394108201360383e-08,
      "loss": 0.4035,
      "num_input_tokens_seen": 24396328,
      "step": 37320
    },
    {
      "epoch": 19.56236897274633,
      "grad_norm": 0.3148244023323059,
      "learning_rate": 7.306491851529206e-08,
      "loss": 0.3631,
      "num_input_tokens_seen": 24404040,
      "step": 37325
    },
    {
      "epoch": 19.564989517819708,
      "grad_norm": 0.5940560102462769,
      "learning_rate": 7.219396944741774e-08,
      "loss": 0.4798,
      "num_input_tokens_seen": 24407304,
      "step": 37330
    },
    {
      "epoch": 19.56761006289308,
      "grad_norm": 0.38467729091644287,
      "learning_rate": 7.132823499217123e-08,
      "loss": 0.5418,
      "num_input_tokens_seen": 24411496,
      "step": 37335
    },
    {
      "epoch": 19.570230607966458,
      "grad_norm": 0.8241835832595825,
      "learning_rate": 7.046771533065766e-08,
      "loss": 0.4115,
      "num_input_tokens_seen": 24414280,
      "step": 37340
    },
    {
      "epoch": 19.57285115303983,
      "grad_norm": 0.45491260290145874,
      "learning_rate": 6.961241064288859e-08,
      "loss": 0.3438,
      "num_input_tokens_seen": 24417352,
      "step": 37345
    },
    {
      "epoch": 19.57547169811321,
      "grad_norm": 0.9295822381973267,
      "learning_rate": 6.876232110778202e-08,
      "loss": 0.5177,
      "num_input_tokens_seen": 24421192,
      "step": 37350
    },
    {
      "epoch": 19.578092243186582,
      "grad_norm": 0.48403412103652954,
      "learning_rate": 6.791744690316515e-08,
      "loss": 0.4716,
      "num_input_tokens_seen": 24424744,
      "step": 37355
    },
    {
      "epoch": 19.58071278825996,
      "grad_norm": 0.44026798009872437,
      "learning_rate": 6.707778820577715e-08,
      "loss": 0.4328,
      "num_input_tokens_seen": 24427592,
      "step": 37360
    },
    {
      "epoch": 19.583333333333332,
      "grad_norm": 0.5145461559295654,
      "learning_rate": 6.624334519126919e-08,
      "loss": 0.5384,
      "num_input_tokens_seen": 24430600,
      "step": 37365
    },
    {
      "epoch": 19.58595387840671,
      "grad_norm": 0.45401260256767273,
      "learning_rate": 6.541411803419329e-08,
      "loss": 0.4324,
      "num_input_tokens_seen": 24433160,
      "step": 37370
    },
    {
      "epoch": 19.588574423480082,
      "grad_norm": 0.6941284537315369,
      "learning_rate": 6.459010690801626e-08,
      "loss": 0.5886,
      "num_input_tokens_seen": 24435656,
      "step": 37375
    },
    {
      "epoch": 19.59119496855346,
      "grad_norm": 0.7413131594657898,
      "learning_rate": 6.377131198510855e-08,
      "loss": 0.5076,
      "num_input_tokens_seen": 24438440,
      "step": 37380
    },
    {
      "epoch": 19.593815513626833,
      "grad_norm": 0.4281684160232544,
      "learning_rate": 6.295773343675537e-08,
      "loss": 0.4128,
      "num_input_tokens_seen": 24441416,
      "step": 37385
    },
    {
      "epoch": 19.59643605870021,
      "grad_norm": 0.46266499161720276,
      "learning_rate": 6.214937143314837e-08,
      "loss": 0.3427,
      "num_input_tokens_seen": 24444328,
      "step": 37390
    },
    {
      "epoch": 19.599056603773583,
      "grad_norm": 1.6010653972625732,
      "learning_rate": 6.134622614338559e-08,
      "loss": 0.4014,
      "num_input_tokens_seen": 24446472,
      "step": 37395
    },
    {
      "epoch": 19.60167714884696,
      "grad_norm": 0.7628206014633179,
      "learning_rate": 6.05482977354771e-08,
      "loss": 0.4751,
      "num_input_tokens_seen": 24449544,
      "step": 37400
    },
    {
      "epoch": 19.604297693920337,
      "grad_norm": 0.4536610543727875,
      "learning_rate": 5.975558637634215e-08,
      "loss": 0.4578,
      "num_input_tokens_seen": 24452776,
      "step": 37405
    },
    {
      "epoch": 19.60691823899371,
      "grad_norm": 0.36305227875709534,
      "learning_rate": 5.896809223180366e-08,
      "loss": 0.4429,
      "num_input_tokens_seen": 24455560,
      "step": 37410
    },
    {
      "epoch": 19.609538784067087,
      "grad_norm": 0.9027189612388611,
      "learning_rate": 5.8185815466596516e-08,
      "loss": 0.4642,
      "num_input_tokens_seen": 24458408,
      "step": 37415
    },
    {
      "epoch": 19.61215932914046,
      "grad_norm": 0.5067326426506042,
      "learning_rate": 5.7408756244370365e-08,
      "loss": 0.3711,
      "num_input_tokens_seen": 24461576,
      "step": 37420
    },
    {
      "epoch": 19.614779874213838,
      "grad_norm": 0.37800443172454834,
      "learning_rate": 5.663691472766741e-08,
      "loss": 0.4304,
      "num_input_tokens_seen": 24464840,
      "step": 37425
    },
    {
      "epoch": 19.61740041928721,
      "grad_norm": 0.4880770742893219,
      "learning_rate": 5.587029107795572e-08,
      "loss": 0.4123,
      "num_input_tokens_seen": 24468200,
      "step": 37430
    },
    {
      "epoch": 19.620020964360588,
      "grad_norm": 0.5660058856010437,
      "learning_rate": 5.510888545559867e-08,
      "loss": 0.4981,
      "num_input_tokens_seen": 24471176,
      "step": 37435
    },
    {
      "epoch": 19.62264150943396,
      "grad_norm": 0.40496712923049927,
      "learning_rate": 5.435269801987997e-08,
      "loss": 0.3239,
      "num_input_tokens_seen": 24473960,
      "step": 37440
    },
    {
      "epoch": 19.62526205450734,
      "grad_norm": 0.5746840238571167,
      "learning_rate": 5.360172892897863e-08,
      "loss": 0.341,
      "num_input_tokens_seen": 24476776,
      "step": 37445
    },
    {
      "epoch": 19.627882599580712,
      "grad_norm": 0.28984785079956055,
      "learning_rate": 5.285597833999401e-08,
      "loss": 0.483,
      "num_input_tokens_seen": 24480456,
      "step": 37450
    },
    {
      "epoch": 19.63050314465409,
      "grad_norm": 0.6712492108345032,
      "learning_rate": 5.2115446408926295e-08,
      "loss": 0.3244,
      "num_input_tokens_seen": 24483400,
      "step": 37455
    },
    {
      "epoch": 19.633123689727462,
      "grad_norm": 0.4911464750766754,
      "learning_rate": 5.138013329068769e-08,
      "loss": 0.5697,
      "num_input_tokens_seen": 24486536,
      "step": 37460
    },
    {
      "epoch": 19.63574423480084,
      "grad_norm": 0.31634098291397095,
      "learning_rate": 5.065003913909405e-08,
      "loss": 0.3423,
      "num_input_tokens_seen": 24490056,
      "step": 37465
    },
    {
      "epoch": 19.638364779874212,
      "grad_norm": 0.47263920307159424,
      "learning_rate": 4.9925164106875975e-08,
      "loss": 0.4421,
      "num_input_tokens_seen": 24493160,
      "step": 37470
    },
    {
      "epoch": 19.64098532494759,
      "grad_norm": 0.7314928770065308,
      "learning_rate": 4.9205508345670506e-08,
      "loss": 0.5037,
      "num_input_tokens_seen": 24496200,
      "step": 37475
    },
    {
      "epoch": 19.643605870020963,
      "grad_norm": 0.45672452449798584,
      "learning_rate": 4.8491072006015546e-08,
      "loss": 0.3433,
      "num_input_tokens_seen": 24500680,
      "step": 37480
    },
    {
      "epoch": 19.64622641509434,
      "grad_norm": 0.4798952341079712,
      "learning_rate": 4.778185523736933e-08,
      "loss": 0.6433,
      "num_input_tokens_seen": 24503720,
      "step": 37485
    },
    {
      "epoch": 19.648846960167717,
      "grad_norm": 0.4737417995929718,
      "learning_rate": 4.7077858188085414e-08,
      "loss": 0.4494,
      "num_input_tokens_seen": 24506440,
      "step": 37490
    },
    {
      "epoch": 19.65146750524109,
      "grad_norm": 0.5106804966926575,
      "learning_rate": 4.6379081005437644e-08,
      "loss": 0.5013,
      "num_input_tokens_seen": 24509384,
      "step": 37495
    },
    {
      "epoch": 19.654088050314467,
      "grad_norm": 0.5211506485939026,
      "learning_rate": 4.568552383560076e-08,
      "loss": 0.4148,
      "num_input_tokens_seen": 24511720,
      "step": 37500
    },
    {
      "epoch": 19.65670859538784,
      "grad_norm": 0.43571290373802185,
      "learning_rate": 4.4997186823655945e-08,
      "loss": 0.4663,
      "num_input_tokens_seen": 24514472,
      "step": 37505
    },
    {
      "epoch": 19.659329140461217,
      "grad_norm": 0.30517059564590454,
      "learning_rate": 4.431407011360189e-08,
      "loss": 0.4162,
      "num_input_tokens_seen": 24517224,
      "step": 37510
    },
    {
      "epoch": 19.66194968553459,
      "grad_norm": 0.44306066632270813,
      "learning_rate": 4.3636173848332627e-08,
      "loss": 0.4348,
      "num_input_tokens_seen": 24519944,
      "step": 37515
    },
    {
      "epoch": 19.664570230607968,
      "grad_norm": 0.5934711694717407,
      "learning_rate": 4.2963498169656946e-08,
      "loss": 0.4855,
      "num_input_tokens_seen": 24524840,
      "step": 37520
    },
    {
      "epoch": 19.66719077568134,
      "grad_norm": 0.3497467041015625,
      "learning_rate": 4.229604321829561e-08,
      "loss": 0.4447,
      "num_input_tokens_seen": 24527912,
      "step": 37525
    },
    {
      "epoch": 19.669811320754718,
      "grad_norm": 0.4800508916378021,
      "learning_rate": 4.16338091338675e-08,
      "loss": 0.3428,
      "num_input_tokens_seen": 24530696,
      "step": 37530
    },
    {
      "epoch": 19.67243186582809,
      "grad_norm": 0.4430038630962372,
      "learning_rate": 4.097679605490623e-08,
      "loss": 0.2957,
      "num_input_tokens_seen": 24533256,
      "step": 37535
    },
    {
      "epoch": 19.67505241090147,
      "grad_norm": 0.4441662132740021,
      "learning_rate": 4.032500411885465e-08,
      "loss": 0.487,
      "num_input_tokens_seen": 24536552,
      "step": 37540
    },
    {
      "epoch": 19.677672955974842,
      "grad_norm": 0.5150315165519714,
      "learning_rate": 3.967843346205369e-08,
      "loss": 0.44,
      "num_input_tokens_seen": 24539464,
      "step": 37545
    },
    {
      "epoch": 19.68029350104822,
      "grad_norm": 0.5167524814605713,
      "learning_rate": 3.903708421976182e-08,
      "loss": 0.3768,
      "num_input_tokens_seen": 24542312,
      "step": 37550
    },
    {
      "epoch": 19.682914046121592,
      "grad_norm": 0.8312040567398071,
      "learning_rate": 3.840095652614395e-08,
      "loss": 0.3896,
      "num_input_tokens_seen": 24546056,
      "step": 37555
    },
    {
      "epoch": 19.68553459119497,
      "grad_norm": 0.33464035391807556,
      "learning_rate": 3.777005051426863e-08,
      "loss": 0.4339,
      "num_input_tokens_seen": 24549512,
      "step": 37560
    },
    {
      "epoch": 19.688155136268342,
      "grad_norm": 0.6627606153488159,
      "learning_rate": 3.71443663161164e-08,
      "loss": 0.4094,
      "num_input_tokens_seen": 24553000,
      "step": 37565
    },
    {
      "epoch": 19.69077568134172,
      "grad_norm": 0.374198853969574,
      "learning_rate": 3.652390406256867e-08,
      "loss": 0.658,
      "num_input_tokens_seen": 24555944,
      "step": 37570
    },
    {
      "epoch": 19.693396226415093,
      "grad_norm": 0.6055225729942322,
      "learning_rate": 3.5908663883421624e-08,
      "loss": 0.3453,
      "num_input_tokens_seen": 24558600,
      "step": 37575
    },
    {
      "epoch": 19.69601677148847,
      "grad_norm": 0.49439242482185364,
      "learning_rate": 3.529864590737786e-08,
      "loss": 0.4103,
      "num_input_tokens_seen": 24562440,
      "step": 37580
    },
    {
      "epoch": 19.698637316561843,
      "grad_norm": 0.3281373679637909,
      "learning_rate": 3.4693850262046415e-08,
      "loss": 0.4311,
      "num_input_tokens_seen": 24565736,
      "step": 37585
    },
    {
      "epoch": 19.70125786163522,
      "grad_norm": 0.40861618518829346,
      "learning_rate": 3.409427707394275e-08,
      "loss": 0.4263,
      "num_input_tokens_seen": 24569064,
      "step": 37590
    },
    {
      "epoch": 19.703878406708597,
      "grad_norm": 0.48570749163627625,
      "learning_rate": 3.349992646848599e-08,
      "loss": 0.4361,
      "num_input_tokens_seen": 24572584,
      "step": 37595
    },
    {
      "epoch": 19.70649895178197,
      "grad_norm": 0.7063657641410828,
      "learning_rate": 3.291079857001556e-08,
      "loss": 0.5914,
      "num_input_tokens_seen": 24575240,
      "step": 37600
    },
    {
      "epoch": 19.709119496855347,
      "grad_norm": 0.4636692702770233,
      "learning_rate": 3.2326893501763436e-08,
      "loss": 0.6015,
      "num_input_tokens_seen": 24578408,
      "step": 37605
    },
    {
      "epoch": 19.71174004192872,
      "grad_norm": 0.6563592553138733,
      "learning_rate": 3.174821138588191e-08,
      "loss": 0.407,
      "num_input_tokens_seen": 24581064,
      "step": 37610
    },
    {
      "epoch": 19.714360587002098,
      "grad_norm": 0.4704841375350952,
      "learning_rate": 3.11747523434186e-08,
      "loss": 0.4045,
      "num_input_tokens_seen": 24584200,
      "step": 37615
    },
    {
      "epoch": 19.71698113207547,
      "grad_norm": 0.4994450807571411,
      "learning_rate": 3.060651649434143e-08,
      "loss": 0.5271,
      "num_input_tokens_seen": 24587400,
      "step": 37620
    },
    {
      "epoch": 19.719601677148848,
      "grad_norm": 0.46620887517929077,
      "learning_rate": 3.004350395751088e-08,
      "loss": 0.3444,
      "num_input_tokens_seen": 24590088,
      "step": 37625
    },
    {
      "epoch": 19.72222222222222,
      "grad_norm": 0.7306334972381592,
      "learning_rate": 2.94857148507105e-08,
      "loss": 0.5675,
      "num_input_tokens_seen": 24593064,
      "step": 37630
    },
    {
      "epoch": 19.7248427672956,
      "grad_norm": 0.3305153250694275,
      "learning_rate": 2.893314929061641e-08,
      "loss": 0.4497,
      "num_input_tokens_seen": 24595816,
      "step": 37635
    },
    {
      "epoch": 19.72746331236897,
      "grad_norm": 0.4094527065753937,
      "learning_rate": 2.8385807392825037e-08,
      "loss": 0.3495,
      "num_input_tokens_seen": 24598824,
      "step": 37640
    },
    {
      "epoch": 19.73008385744235,
      "grad_norm": 0.5631872415542603,
      "learning_rate": 2.7843689271830897e-08,
      "loss": 0.4608,
      "num_input_tokens_seen": 24601480,
      "step": 37645
    },
    {
      "epoch": 19.732704402515722,
      "grad_norm": 0.6351056694984436,
      "learning_rate": 2.7306795041040502e-08,
      "loss": 0.358,
      "num_input_tokens_seen": 24604296,
      "step": 37650
    },
    {
      "epoch": 19.7353249475891,
      "grad_norm": 0.2942449450492859,
      "learning_rate": 2.677512481276401e-08,
      "loss": 0.3801,
      "num_input_tokens_seen": 24607336,
      "step": 37655
    },
    {
      "epoch": 19.737945492662472,
      "grad_norm": 0.32324135303497314,
      "learning_rate": 2.6248678698220785e-08,
      "loss": 0.5249,
      "num_input_tokens_seen": 24610984,
      "step": 37660
    },
    {
      "epoch": 19.74056603773585,
      "grad_norm": 0.32838496565818787,
      "learning_rate": 2.5727456807542183e-08,
      "loss": 0.3792,
      "num_input_tokens_seen": 24613800,
      "step": 37665
    },
    {
      "epoch": 19.743186582809223,
      "grad_norm": 0.26942241191864014,
      "learning_rate": 2.5211459249754875e-08,
      "loss": 0.3587,
      "num_input_tokens_seen": 24617064,
      "step": 37670
    },
    {
      "epoch": 19.7458071278826,
      "grad_norm": 0.3557201027870178,
      "learning_rate": 2.4700686132803076e-08,
      "loss": 0.4091,
      "num_input_tokens_seen": 24621416,
      "step": 37675
    },
    {
      "epoch": 19.748427672955973,
      "grad_norm": 0.40408432483673096,
      "learning_rate": 2.4195137563534643e-08,
      "loss": 0.4602,
      "num_input_tokens_seen": 24625160,
      "step": 37680
    },
    {
      "epoch": 19.75104821802935,
      "grad_norm": 0.4115641117095947,
      "learning_rate": 2.369481364770387e-08,
      "loss": 0.3958,
      "num_input_tokens_seen": 24628520,
      "step": 37685
    },
    {
      "epoch": 19.753668763102727,
      "grad_norm": 0.5043158531188965,
      "learning_rate": 2.3199714489974266e-08,
      "loss": 0.5234,
      "num_input_tokens_seen": 24631528,
      "step": 37690
    },
    {
      "epoch": 19.7562893081761,
      "grad_norm": 0.3579798936843872,
      "learning_rate": 2.2709840193912978e-08,
      "loss": 0.5287,
      "num_input_tokens_seen": 24634440,
      "step": 37695
    },
    {
      "epoch": 19.758909853249477,
      "grad_norm": 0.3652504086494446,
      "learning_rate": 2.2225190861999146e-08,
      "loss": 0.4445,
      "num_input_tokens_seen": 24637704,
      "step": 37700
    },
    {
      "epoch": 19.76153039832285,
      "grad_norm": 0.674635112285614,
      "learning_rate": 2.1745766595612783e-08,
      "loss": 0.5668,
      "num_input_tokens_seen": 24640776,
      "step": 37705
    },
    {
      "epoch": 19.764150943396228,
      "grad_norm": 0.5477362275123596,
      "learning_rate": 2.1271567495045886e-08,
      "loss": 0.3072,
      "num_input_tokens_seen": 24643624,
      "step": 37710
    },
    {
      "epoch": 19.7667714884696,
      "grad_norm": 0.398446261882782,
      "learning_rate": 2.0802593659491333e-08,
      "loss": 0.5016,
      "num_input_tokens_seen": 24647496,
      "step": 37715
    },
    {
      "epoch": 19.769392033542978,
      "grad_norm": 0.38867461681365967,
      "learning_rate": 2.0338845187059528e-08,
      "loss": 0.5639,
      "num_input_tokens_seen": 24651048,
      "step": 37720
    },
    {
      "epoch": 19.77201257861635,
      "grad_norm": 0.5195918679237366,
      "learning_rate": 1.988032217475899e-08,
      "loss": 0.4969,
      "num_input_tokens_seen": 24653736,
      "step": 37725
    },
    {
      "epoch": 19.77463312368973,
      "grad_norm": 0.533473014831543,
      "learning_rate": 1.942702471850466e-08,
      "loss": 0.4174,
      "num_input_tokens_seen": 24657032,
      "step": 37730
    },
    {
      "epoch": 19.7772536687631,
      "grad_norm": 0.4316028356552124,
      "learning_rate": 1.897895291312346e-08,
      "loss": 0.4258,
      "num_input_tokens_seen": 24660136,
      "step": 37735
    },
    {
      "epoch": 19.77987421383648,
      "grad_norm": 0.48570847511291504,
      "learning_rate": 1.8536106852348746e-08,
      "loss": 0.5925,
      "num_input_tokens_seen": 24663048,
      "step": 37740
    },
    {
      "epoch": 19.782494758909852,
      "grad_norm": 0.3607500493526459,
      "learning_rate": 1.8098486628814748e-08,
      "loss": 0.4161,
      "num_input_tokens_seen": 24667016,
      "step": 37745
    },
    {
      "epoch": 19.78511530398323,
      "grad_norm": 0.390424907207489,
      "learning_rate": 1.7666092334067684e-08,
      "loss": 0.4694,
      "num_input_tokens_seen": 24670568,
      "step": 37750
    },
    {
      "epoch": 19.787735849056602,
      "grad_norm": 0.6404188275337219,
      "learning_rate": 1.7238924058562977e-08,
      "loss": 0.5188,
      "num_input_tokens_seen": 24674024,
      "step": 37755
    },
    {
      "epoch": 19.79035639412998,
      "grad_norm": 0.6621967554092407,
      "learning_rate": 1.6816981891656924e-08,
      "loss": 0.5386,
      "num_input_tokens_seen": 24677320,
      "step": 37760
    },
    {
      "epoch": 19.792976939203353,
      "grad_norm": 0.41409018635749817,
      "learning_rate": 1.6400265921615032e-08,
      "loss": 0.3922,
      "num_input_tokens_seen": 24681096,
      "step": 37765
    },
    {
      "epoch": 19.79559748427673,
      "grad_norm": 0.35429733991622925,
      "learning_rate": 1.598877623560924e-08,
      "loss": 0.4556,
      "num_input_tokens_seen": 24684168,
      "step": 37770
    },
    {
      "epoch": 19.798218029350103,
      "grad_norm": 0.32437363266944885,
      "learning_rate": 1.5582512919717908e-08,
      "loss": 0.433,
      "num_input_tokens_seen": 24687464,
      "step": 37775
    },
    {
      "epoch": 19.80083857442348,
      "grad_norm": 0.5069809556007385,
      "learning_rate": 1.5181476058928613e-08,
      "loss": 0.4935,
      "num_input_tokens_seen": 24690920,
      "step": 37780
    },
    {
      "epoch": 19.803459119496857,
      "grad_norm": 0.4693477749824524,
      "learning_rate": 1.4785665737129805e-08,
      "loss": 0.3051,
      "num_input_tokens_seen": 24693768,
      "step": 37785
    },
    {
      "epoch": 19.80607966457023,
      "grad_norm": 0.8105985522270203,
      "learning_rate": 1.4395082037127471e-08,
      "loss": 0.4803,
      "num_input_tokens_seen": 24696552,
      "step": 37790
    },
    {
      "epoch": 19.808700209643607,
      "grad_norm": 0.5855545997619629,
      "learning_rate": 1.4009725040620148e-08,
      "loss": 0.3834,
      "num_input_tokens_seen": 24699656,
      "step": 37795
    },
    {
      "epoch": 19.81132075471698,
      "grad_norm": 0.42504677176475525,
      "learning_rate": 1.3629594828223902e-08,
      "loss": 0.5348,
      "num_input_tokens_seen": 24703496,
      "step": 37800
    },
    {
      "epoch": 19.813941299790358,
      "grad_norm": 0.3962383270263672,
      "learning_rate": 1.3254691479458458e-08,
      "loss": 0.5285,
      "num_input_tokens_seen": 24706760,
      "step": 37805
    },
    {
      "epoch": 19.81656184486373,
      "grad_norm": 0.4927203059196472,
      "learning_rate": 1.2885015072747197e-08,
      "loss": 0.4829,
      "num_input_tokens_seen": 24710952,
      "step": 37810
    },
    {
      "epoch": 19.819182389937108,
      "grad_norm": 0.48116806149482727,
      "learning_rate": 1.2520565685419927e-08,
      "loss": 0.32,
      "num_input_tokens_seen": 24714088,
      "step": 37815
    },
    {
      "epoch": 19.82180293501048,
      "grad_norm": 1.2512212991714478,
      "learning_rate": 1.2161343393721214e-08,
      "loss": 0.4383,
      "num_input_tokens_seen": 24717064,
      "step": 37820
    },
    {
      "epoch": 19.82442348008386,
      "grad_norm": 1.1289942264556885,
      "learning_rate": 1.1807348272790952e-08,
      "loss": 0.4915,
      "num_input_tokens_seen": 24720712,
      "step": 37825
    },
    {
      "epoch": 19.82704402515723,
      "grad_norm": 0.6333639025688171,
      "learning_rate": 1.1458580396683794e-08,
      "loss": 0.6143,
      "num_input_tokens_seen": 24723560,
      "step": 37830
    },
    {
      "epoch": 19.82966457023061,
      "grad_norm": 0.3522340655326843,
      "learning_rate": 1.1115039838358044e-08,
      "loss": 0.5707,
      "num_input_tokens_seen": 24731656,
      "step": 37835
    },
    {
      "epoch": 19.832285115303982,
      "grad_norm": 0.8651641607284546,
      "learning_rate": 1.0776726669678439e-08,
      "loss": 0.6379,
      "num_input_tokens_seen": 24734280,
      "step": 37840
    },
    {
      "epoch": 19.83490566037736,
      "grad_norm": 1.6565288305282593,
      "learning_rate": 1.044364096141337e-08,
      "loss": 0.5822,
      "num_input_tokens_seen": 24736840,
      "step": 37845
    },
    {
      "epoch": 19.837526205450732,
      "grad_norm": 0.2647029459476471,
      "learning_rate": 1.0115782783243211e-08,
      "loss": 0.4223,
      "num_input_tokens_seen": 24740200,
      "step": 37850
    },
    {
      "epoch": 19.84014675052411,
      "grad_norm": 0.4234452545642853,
      "learning_rate": 9.793152203751988e-09,
      "loss": 0.402,
      "num_input_tokens_seen": 24743592,
      "step": 37855
    },
    {
      "epoch": 19.842767295597483,
      "grad_norm": 0.8115110397338867,
      "learning_rate": 9.475749290430158e-09,
      "loss": 0.4387,
      "num_input_tokens_seen": 24746344,
      "step": 37860
    },
    {
      "epoch": 19.84538784067086,
      "grad_norm": 0.5066845417022705,
      "learning_rate": 9.16357410967461e-09,
      "loss": 0.3425,
      "num_input_tokens_seen": 24748712,
      "step": 37865
    },
    {
      "epoch": 19.848008385744233,
      "grad_norm": 0.4936116337776184,
      "learning_rate": 8.856626726788664e-09,
      "loss": 0.431,
      "num_input_tokens_seen": 24752264,
      "step": 37870
    },
    {
      "epoch": 19.85062893081761,
      "grad_norm": 0.47131380438804626,
      "learning_rate": 8.554907205982066e-09,
      "loss": 0.3426,
      "num_input_tokens_seen": 24755816,
      "step": 37875
    },
    {
      "epoch": 19.853249475890987,
      "grad_norm": 0.6134276986122131,
      "learning_rate": 8.258415610373772e-09,
      "loss": 0.4623,
      "num_input_tokens_seen": 24758792,
      "step": 37880
    },
    {
      "epoch": 19.85587002096436,
      "grad_norm": 0.37095677852630615,
      "learning_rate": 7.967152001980838e-09,
      "loss": 0.4056,
      "num_input_tokens_seen": 24762504,
      "step": 37885
    },
    {
      "epoch": 19.858490566037737,
      "grad_norm": 0.5208348035812378,
      "learning_rate": 7.681116441737856e-09,
      "loss": 0.5382,
      "num_input_tokens_seen": 24766760,
      "step": 37890
    },
    {
      "epoch": 19.86111111111111,
      "grad_norm": 0.6228058934211731,
      "learning_rate": 7.4003089894747464e-09,
      "loss": 0.4877,
      "num_input_tokens_seen": 24770536,
      "step": 37895
    },
    {
      "epoch": 19.863731656184488,
      "grad_norm": 0.48377013206481934,
      "learning_rate": 7.12472970393896e-09,
      "loss": 0.3968,
      "num_input_tokens_seen": 24773960,
      "step": 37900
    },
    {
      "epoch": 19.86635220125786,
      "grad_norm": 0.4638943672180176,
      "learning_rate": 6.854378642776049e-09,
      "loss": 0.3865,
      "num_input_tokens_seen": 24776744,
      "step": 37905
    },
    {
      "epoch": 19.868972746331238,
      "grad_norm": 0.35095635056495667,
      "learning_rate": 6.589255862538002e-09,
      "loss": 0.4728,
      "num_input_tokens_seen": 24781128,
      "step": 37910
    },
    {
      "epoch": 19.87159329140461,
      "grad_norm": 0.5017642974853516,
      "learning_rate": 6.32936141869156e-09,
      "loss": 0.3655,
      "num_input_tokens_seen": 24784072,
      "step": 37915
    },
    {
      "epoch": 19.87421383647799,
      "grad_norm": 1.0077147483825684,
      "learning_rate": 6.074695365596017e-09,
      "loss": 0.4791,
      "num_input_tokens_seen": 24787304,
      "step": 37920
    },
    {
      "epoch": 19.87683438155136,
      "grad_norm": 1.023657202720642,
      "learning_rate": 5.8252577565309775e-09,
      "loss": 0.3939,
      "num_input_tokens_seen": 24790216,
      "step": 37925
    },
    {
      "epoch": 19.87945492662474,
      "grad_norm": 0.3863258957862854,
      "learning_rate": 5.581048643671371e-09,
      "loss": 0.4655,
      "num_input_tokens_seen": 24794024,
      "step": 37930
    },
    {
      "epoch": 19.882075471698112,
      "grad_norm": 0.7351384162902832,
      "learning_rate": 5.342068078106888e-09,
      "loss": 0.4257,
      "num_input_tokens_seen": 24797032,
      "step": 37935
    },
    {
      "epoch": 19.88469601677149,
      "grad_norm": 0.5603712201118469,
      "learning_rate": 5.1083161098253175e-09,
      "loss": 0.4,
      "num_input_tokens_seen": 24799592,
      "step": 37940
    },
    {
      "epoch": 19.887316561844862,
      "grad_norm": 0.7794079184532166,
      "learning_rate": 4.87979278772921e-09,
      "loss": 0.4392,
      "num_input_tokens_seen": 24803944,
      "step": 37945
    },
    {
      "epoch": 19.88993710691824,
      "grad_norm": 0.42481330037117004,
      "learning_rate": 4.656498159619216e-09,
      "loss": 0.4957,
      "num_input_tokens_seen": 24807816,
      "step": 37950
    },
    {
      "epoch": 19.892557651991613,
      "grad_norm": 0.7121046185493469,
      "learning_rate": 4.438432272210746e-09,
      "loss": 0.3922,
      "num_input_tokens_seen": 24811272,
      "step": 37955
    },
    {
      "epoch": 19.89517819706499,
      "grad_norm": 0.4501437842845917,
      "learning_rate": 4.225595171114538e-09,
      "loss": 0.4115,
      "num_input_tokens_seen": 24815048,
      "step": 37960
    },
    {
      "epoch": 19.897798742138363,
      "grad_norm": 0.3749341070652008,
      "learning_rate": 4.017986900858861e-09,
      "loss": 0.4815,
      "num_input_tokens_seen": 24818120,
      "step": 37965
    },
    {
      "epoch": 19.90041928721174,
      "grad_norm": 0.6632663607597351,
      "learning_rate": 3.815607504870089e-09,
      "loss": 0.4264,
      "num_input_tokens_seen": 24822280,
      "step": 37970
    },
    {
      "epoch": 19.903039832285117,
      "grad_norm": 0.3091784119606018,
      "learning_rate": 3.618457025486577e-09,
      "loss": 0.4423,
      "num_input_tokens_seen": 24826568,
      "step": 37975
    },
    {
      "epoch": 19.90566037735849,
      "grad_norm": 0.34100908041000366,
      "learning_rate": 3.426535503947559e-09,
      "loss": 0.4445,
      "num_input_tokens_seen": 24829768,
      "step": 37980
    },
    {
      "epoch": 19.908280922431867,
      "grad_norm": 1.3371878862380981,
      "learning_rate": 3.2398429803986997e-09,
      "loss": 0.4009,
      "num_input_tokens_seen": 24832392,
      "step": 37985
    },
    {
      "epoch": 19.91090146750524,
      "grad_norm": 0.8002299070358276,
      "learning_rate": 3.058379493900421e-09,
      "loss": 0.5549,
      "num_input_tokens_seen": 24834856,
      "step": 37990
    },
    {
      "epoch": 19.913522012578618,
      "grad_norm": 0.37803202867507935,
      "learning_rate": 2.882145082405696e-09,
      "loss": 0.606,
      "num_input_tokens_seen": 24838248,
      "step": 37995
    },
    {
      "epoch": 19.91614255765199,
      "grad_norm": 0.6127198934555054,
      "learning_rate": 2.711139782785033e-09,
      "loss": 0.3747,
      "num_input_tokens_seen": 24841864,
      "step": 38000
    },
    {
      "epoch": 19.918763102725368,
      "grad_norm": 0.7399004697799683,
      "learning_rate": 2.5453636308098163e-09,
      "loss": 0.4352,
      "num_input_tokens_seen": 24844776,
      "step": 38005
    },
    {
      "epoch": 19.92138364779874,
      "grad_norm": 0.5057812333106995,
      "learning_rate": 2.3848166611606383e-09,
      "loss": 0.4518,
      "num_input_tokens_seen": 24848360,
      "step": 38010
    },
    {
      "epoch": 19.92400419287212,
      "grad_norm": 0.48696503043174744,
      "learning_rate": 2.2294989074161943e-09,
      "loss": 0.4942,
      "num_input_tokens_seen": 24851496,
      "step": 38015
    },
    {
      "epoch": 19.92662473794549,
      "grad_norm": 0.411210834980011,
      "learning_rate": 2.0794104020727125e-09,
      "loss": 0.379,
      "num_input_tokens_seen": 24855144,
      "step": 38020
    },
    {
      "epoch": 19.92924528301887,
      "grad_norm": 0.4886453151702881,
      "learning_rate": 1.9345511765272993e-09,
      "loss": 0.4124,
      "num_input_tokens_seen": 24858120,
      "step": 38025
    },
    {
      "epoch": 19.931865828092242,
      "grad_norm": 0.5265988111495972,
      "learning_rate": 1.794921261077942e-09,
      "loss": 0.4846,
      "num_input_tokens_seen": 24862248,
      "step": 38030
    },
    {
      "epoch": 19.93448637316562,
      "grad_norm": 0.24993127584457397,
      "learning_rate": 1.6605206849373833e-09,
      "loss": 0.4611,
      "num_input_tokens_seen": 24864776,
      "step": 38035
    },
    {
      "epoch": 19.937106918238992,
      "grad_norm": 0.3758941888809204,
      "learning_rate": 1.531349476219246e-09,
      "loss": 0.4588,
      "num_input_tokens_seen": 24868200,
      "step": 38040
    },
    {
      "epoch": 19.93972746331237,
      "grad_norm": 0.42801472544670105,
      "learning_rate": 1.4074076619463584e-09,
      "loss": 0.4189,
      "num_input_tokens_seen": 24870824,
      "step": 38045
    },
    {
      "epoch": 19.942348008385743,
      "grad_norm": 0.5100420713424683,
      "learning_rate": 1.2886952680452036e-09,
      "loss": 0.4505,
      "num_input_tokens_seen": 24874056,
      "step": 38050
    },
    {
      "epoch": 19.94496855345912,
      "grad_norm": 0.9123054146766663,
      "learning_rate": 1.1752123193459197e-09,
      "loss": 0.3713,
      "num_input_tokens_seen": 24877096,
      "step": 38055
    },
    {
      "epoch": 19.947589098532493,
      "grad_norm": 0.492792546749115,
      "learning_rate": 1.0669588395934017e-09,
      "loss": 0.4185,
      "num_input_tokens_seen": 24880104,
      "step": 38060
    },
    {
      "epoch": 19.95020964360587,
      "grad_norm": 0.48596611618995667,
      "learning_rate": 9.639348514278722e-10,
      "loss": 0.5079,
      "num_input_tokens_seen": 24883112,
      "step": 38065
    },
    {
      "epoch": 19.952830188679247,
      "grad_norm": 0.5100896954536438,
      "learning_rate": 8.661403764043119e-10,
      "loss": 0.4459,
      "num_input_tokens_seen": 24886824,
      "step": 38070
    },
    {
      "epoch": 19.95545073375262,
      "grad_norm": 0.763964831829071,
      "learning_rate": 7.735754349785795e-10,
      "loss": 0.4265,
      "num_input_tokens_seen": 24889096,
      "step": 38075
    },
    {
      "epoch": 19.958071278825997,
      "grad_norm": 0.49122506380081177,
      "learning_rate": 6.862400465157403e-10,
      "loss": 0.41,
      "num_input_tokens_seen": 24893640,
      "step": 38080
    },
    {
      "epoch": 19.96069182389937,
      "grad_norm": 0.4501722455024719,
      "learning_rate": 6.041342292817387e-10,
      "loss": 0.4807,
      "num_input_tokens_seen": 24896264,
      "step": 38085
    },
    {
      "epoch": 19.963312368972748,
      "grad_norm": 0.434285968542099,
      "learning_rate": 5.272580004572758e-10,
      "loss": 0.4063,
      "num_input_tokens_seen": 24899080,
      "step": 38090
    },
    {
      "epoch": 19.96593291404612,
      "grad_norm": 0.5897517204284668,
      "learning_rate": 4.556113761211567e-10,
      "loss": 0.4219,
      "num_input_tokens_seen": 24902024,
      "step": 38095
    },
    {
      "epoch": 19.968553459119498,
      "grad_norm": 0.47378137707710266,
      "learning_rate": 3.8919437125861657e-10,
      "loss": 0.3893,
      "num_input_tokens_seen": 24906088,
      "step": 38100
    },
    {
      "epoch": 19.97117400419287,
      "grad_norm": 1.1681697368621826,
      "learning_rate": 3.280069997696478e-10,
      "loss": 0.4853,
      "num_input_tokens_seen": 24908968,
      "step": 38105
    },
    {
      "epoch": 19.97379454926625,
      "grad_norm": 0.43145594000816345,
      "learning_rate": 2.720492744467951e-10,
      "loss": 0.4606,
      "num_input_tokens_seen": 24912104,
      "step": 38110
    },
    {
      "epoch": 19.97641509433962,
      "grad_norm": 0.6073048710823059,
      "learning_rate": 2.2132120700291136e-10,
      "loss": 0.4902,
      "num_input_tokens_seen": 24915464,
      "step": 38115
    },
    {
      "epoch": 19.979035639413,
      "grad_norm": 0.49991512298583984,
      "learning_rate": 1.7582280804340213e-10,
      "loss": 0.3711,
      "num_input_tokens_seen": 24918632,
      "step": 38120
    },
    {
      "epoch": 19.981656184486372,
      "grad_norm": 0.7059888243675232,
      "learning_rate": 1.3555408708842977e-10,
      "loss": 0.5329,
      "num_input_tokens_seen": 24921032,
      "step": 38125
    },
    {
      "epoch": 19.98427672955975,
      "grad_norm": 0.4821523129940033,
      "learning_rate": 1.0051505256458704e-10,
      "loss": 0.4413,
      "num_input_tokens_seen": 24923912,
      "step": 38130
    },
    {
      "epoch": 19.986897274633122,
      "grad_norm": 0.48262301087379456,
      "learning_rate": 7.070571179657037e-11,
      "loss": 0.4817,
      "num_input_tokens_seen": 24926952,
      "step": 38135
    },
    {
      "epoch": 19.9895178197065,
      "grad_norm": 0.27878475189208984,
      "learning_rate": 4.612607102383315e-11,
      "loss": 0.4958,
      "num_input_tokens_seen": 24930984,
      "step": 38140
    },
    {
      "epoch": 19.992138364779873,
      "grad_norm": 0.47275716066360474,
      "learning_rate": 2.677613538670798e-11,
      "loss": 0.5802,
      "num_input_tokens_seen": 24933832,
      "step": 38145
    },
    {
      "epoch": 19.99475890985325,
      "grad_norm": 0.40565821528434753,
      "learning_rate": 1.2655908931957782e-11,
      "loss": 0.4092,
      "num_input_tokens_seen": 24936584,
      "step": 38150
    },
    {
      "epoch": 19.997379454926623,
      "grad_norm": 1.0477665662765503,
      "learning_rate": 3.765394615551366e-12,
      "loss": 0.362,
      "num_input_tokens_seen": 24939432,
      "step": 38155
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.9950681924819946,
      "learning_rate": 1.0459429711229262e-13,
      "loss": 0.614,
      "num_input_tokens_seen": 24941912,
      "step": 38160
    },
    {
      "epoch": 20.0,
      "eval_loss": 0.4749854803085327,
      "eval_runtime": 9.7216,
      "eval_samples_per_second": 87.229,
      "eval_steps_per_second": 21.807,
      "num_input_tokens_seen": 24941912,
      "step": 38160
    },
    {
      "epoch": 20.0,
      "num_input_tokens_seen": 24941912,
      "step": 38160,
      "total_flos": 1.1231628676218225e+18,
      "train_loss": 0.4822516317668701,
      "train_runtime": 6549.5572,
      "train_samples_per_second": 23.296,
      "train_steps_per_second": 5.826
    }
  ],
  "logging_steps": 5,
  "max_steps": 38160,
  "num_input_tokens_seen": 24941912,
  "num_train_epochs": 20,
  "save_steps": 1908,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1231628676218225e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}